Public Interest Energy Research (PIER) Program INTERIM PROJECT REPORT

HUMBOLDT COUNTY AS A RENEWABLE ENERGY SECURE COMMUNITY

Resource and Technology Assessment Report Appendices - DRAFT

Prepared for: California Energy Commission

Prepared by: Schatz Energy Research Center, Humboldt State University

AUGUST 2012

CEC-XXX-2012-

Prepared by:

Primary Author(s): Jim Zoellick, Senior Research Engineer, SERC Colin Sheppard, Research Engineer, SERC Peter Alstone, Research Engineer, SERC

Also Contributing: Andrea Alstone, Dr. Charles Chamberlin, Ruben Garcia, Dr. Steven Hackett, Dr. Peter Lehman, Eric Martin, Luke Schiedler, SERC

Schatz Energy Research Center Humboldt State University Arcata, CA 95521 707-826-4345 www.schatzlab.org

Contract Number: PIR-08-034

Prepared for: Public Interest Energy Research (PIER) California Energy Commission

Michael Sokol Project Manager

Linda Spiegel Office Manager Energy Generation Research Office

Laurie ten Hope Deputy Director Energy Research & Development Division

Robert P. Oglesby Executive Director

DISCLAIMER

This report was prepared as the result of work sponsored by the California Energy Commission. It does not necessarily represent the views of the Energy Commission, its employees or the State of California. The Energy Commission, the State of California, its employees, contractors and subcontractors make no warrant, express or implied, and assume no legal liability for the information in this report; nor does any party represent that the uses of this information will not infringe upon privately owned rights. This report has not been approved or disapproved by the California Energy Commission nor has the California Energy Commission passed upon the accuracy or adequacy of the information in this report.

Copyright © 2012 Schatz Energy Research Center APPENDIX A: Electric Vehicle Assessment

A.1 Analysis Methodology

A.1.1 Baseline

To assess the impacts that electric vehicles (EVs) could have in Humboldt County it was first necessary to characterize the existing vehicle fleet as a basis for comparison. To establish such a baseline, county level transportation data were collected from the California Department of Motor Vehicles and the California Department of Transportation. Using these data, registered vehicles, annual vehicle miles traveled, and annual vehicle fuel consumption were extrapolated to estimate future growth. National fuel economy estimates by the Bureau of Transportation Statistics were also used to estimate future baseline fleet vehicle fuel consumption in Humboldt County. An annual vehicle fuel consumption growth rate of 0.5 percent was used based on average annual vehicle fuel consumption growth from 1999 to 2006. This suggests that about 76.3 and 80.2 million gallons of fuel would be consumed in 2020 and 2030, respectively. This corresponds to an average of 11,000 miles of travel per vehicle-year. A.1.2 Vehicle Adoption Rate

To project the adoption of electric vehicles in Humboldt County, two primary literature sources were consulted. This included the Environmental Assessment of Plug-In Hybrid Electric Vehicles by EPRI and the Natural Resource Defense Council (EPRI/NRDC, 2007) as well as the Long-Range EV Charging Infrastructure Plan for Western Oregon by the Electric Transportation Engineering Corporation (ETEC, 2010). Integral to the EPRI/NRDC study was the development of market adoption scenarios, which estimate the new vehicle market share of plug-in hybrid electric vehicles from 2010 to 2050. Low, Medium, and High scenarios were developed and achieve a market share of 20 percent, 51 percent, and 68 percent, respectively, by 2030, and a maximum market share of 20 percent, 62 percent, and 80 percent, respectively, by 2050 (Figure 25). In comparison to other projections in the literature, EPRI’s Medium and High plug-in hybrid electric vehicle scenarios are aggressive. The underlying assumptions for these projections are not specifically outlined, though they are intended to apply at the national level. As Humboldt County is a rather isolated rural area in Northwestern California, it is questionable how well these penetration rates may apply. State or regional market share projections may be more applicable.

The Electric Transportation Engineering Corporation study (ETEC, 2010) was also consulted. In August of 2009, ETEC was awarded a $99.8 million grant from the U.S. Department of Energy to partner with Nissan to deploy up to 4,700 battery electric vehicles (BEVs, for example, Nissan Leafs) and 11,210 charging systems in strategic markets in the states of Arizona, California, Oregon, Tennessee, and Washington (ETEC,

87 2010). The ETEC study presents various electric vehicle penetration projections for the U.S. and Western Oregon. Given Humboldt County’s proximity and demographic similarities to the Corvallis-Albany region cited in the ETEC report, electric vehicle market share projections for the Corvallis-Albany area were assumed applicable and scaled to Humboldt County. Figure 25: Assumed New Vehicle Market Share Under a Medium Plug-In Hybrid Electric Vehicle Scenario

Source: EPRI (2007)

Figure 26 shows the projected electric vehicle market share for Corvallis-Albany in comparison to plug-in hybrid electric vehicles market share projected for the U.S. according to the aforementioned EPRI scenarios. The ETEC Corvallis scenario most closely follows the EPRI Low scenario. Doubling the Corvallis scenario penetration rates achieves a 2030 market penetration (53 percent) similar to that of the EPRI Medium scenario (51 percent), but is much slower in reaching that point. Figure 27 shows cumulative electric vehicles in Humboldt County according to the aforementioned electric vehicles market share projections. Three electric vehicle penetration scenarios were assumed for Humboldt County: (1) A likely (medium) scenario corresponding to expected electric vehicle diffusion in Corvallis-Albany; (2) an optimistic (high) scenario corresponding to a doubling of the expected penetration in Corvallis-Albany; and (3) a maximum scenario corresponding to the EPRI Medium scenario, meant to serve as a cap for electric vehicles penetration in the REPOP Model. This resulted in a cap on cumulative vehicle penetration in 2030 of 38 percent.

88 Figure 26: Assumed New Car Market Share in the United States for Low, Medium, and High Plug-in Hybrid Electric Vehicle Scenarios and Corvallis-Albany EV Scenario

70%

60%

50%

40%

30%

20% % of New Vehicle Sales 10%

0% 2010 2015 2020 2025 2030 ETEC: Corvallis ETEC: Corvallis (x2) EPRI: Low EPRI: Medium EPRI: High

Source: SERC compilation from EPRI (2007) and ETEC (2010)

Figure 27: Projected Cumulative EVs in Humboldt County Based on Low, Medium, and High Plug-in Hybrid Electric Vehicle Scenarios and a Corvallis-Albany EV Scenario

90,000 80,000 70,000 60,000 50,000 40,000

Cumulave EVs 30,000 20,000 10,000 0 2010 2015 2020 2025 2030 ETEC: Corvallis ETEC: Corvallis (x2) EPRI: Low EPRI: Medium EPRI: High

Source: SERC compilation from EPRI (2007) and ETEC (2010)

89 A.1.3 Vehicle Performance

Most assumptions about electric vehicle architecture, performance, energy consumption, and capital and maintenance costs were obtained from the literature source Where Are the Market Niches for Electric Drive Vehicles by Santini et al. (2010), as well as via personal communication with Bryan D. Jungers, Research Manager at E Source in Boulder, Colorado. Only five electric vehicles were considered for this analysis based on architectures that are representative of current or near market technologies (Table 15). Santini et al. (2010) used the Powertrain System Analysis Toolkit to model and evaluate the performance of nine electric vehicles spanning the major classes of electric drive vehicles, including the split parallel-and-series system used by Toyota and Ford, a series extended range electric vehicle (EREV) similar to the Chevrolet Volt, and a battery electric powertrain similar to the Nissan Leaf,, Table 16.. Table 15: Description of Conventional and Electric Vehicles Considered for Adoption

Vehicle Description Conventional vehicle; standard mid-size sedan powered by an internal Conventional combustion engine (for example, Toyota Camry) Plug-in hybrid electric vehicle with a power-split drivetrain and an all- PHEV10 electric range of 10 miles (for example, Toyota Plug-in Prius) Plug-in hybrid electric vehicle with a series drivetrain and an all-electric PHEV40 range of 40 miles Extended range electric vehicle; a plug-in hybrid electric vehicle with a series powertrain and an all-electric range of 30 miles; EREVs and series PHEVs are virtually identical, except that an EREV has an over-sized electric motor and battery that enable aggressive acceleration without EREV30 triggering an engine-on event Extended range electric vehicle; a plug-in hybrid electric vehicle with a series powertrain and an all-electric range of 40 miles (for example, the Chevrolet Volt was originally advertised as an EREV40, but is reportedly EREV40 a PHEV40 with a power-split drivetrain) Battery electric vehicle with an all-electric range of 100 miles (for BEV100 example, Nissan Leaf)

Source: SERC staff.

90 Table 16: Drivetrain Components of Vehicles Considered for Adoption Gross MC2/Gen MC2/Gen ICE MC Peak MC Cont. Component Vehicle Peak Cont. Power Power Power Weight Power Power Vehicle Name (kg) (kW) (kW) (kW) (kW) (kW) Conventional 2111 135 0 0 2 2 PHEV10 2201 80 76 38 45 30 PHEV40 2321 70 108 54 68.25 68.25 EREV30 2325 75 109 54.5 73.2 73.2 EREV40 2355 75 109 54.5 73.2 73.2 BEV100 2185 0 103 51.5 0 0 BOL ESS EOL ESS BOL ESS EOL ESS CD Total Total Power-to- Power-to- Degree- Component Useable Nominal Nominal Energy Energy of-Elec. Energy Energy Energy Ratio Ratio Vehicle Name (kWh) (kWh) (1/hr) (1/hr) (%) (kWh) Conventional 0.0 0.0 0.0 0.0 0% 0.00 PHEV10 5.0 4.0 15.5 15.5 41% 2.30 PHEV40 16.9 13.5 4.4 4.4 44% 8.96 EREV30 12.4 9.9 15.4 15.4 59% 6.42 EREV40 16.3 13.1 12.1 12.1 59% 8.64 BEV100 33.2 26.6 5.8 8.3 100% 18.6 Component Definitions ICE: internal combustion engine MC: electric machine/motor MC2: electric machine/motor #2 Gen: generator BOL: beginning of life EOL: end of life ESS: energy storage system CD: charge-depleting

Source: Santini et al. (2010)

Vehicle fuel economy and electric efficiency are shown in Table 17 and include minor improvements in fuel economy expected to occur by 2020. Based on an estimated 60 percent highway and 40 percent city split in vehicle miles traveled in Humboldt County, Figure 28 compares the tank-to-wheel efficiency of electric vehicles in charge-depleting mode and a conventional vehicle. On average, a conventional vehicle requires about four times the energy as an electric vehicle on a per mile basis. This is reasonable since an electric motor is much more efficient than an ICE and highlights one of the major advantages of an EV. Assuming approximate current day gasoline and electricity costs ($4/gallon, $0.12/kWh) in Humboldt County, a conventional vehicle costs about 3.6 times as much to fuel as the average EV (Figure 29). Electric vehicles are clearly more efficient which results in less fuel consumption and fuel costs for vehicle owners.

91 Table 17: Fuel and Electric Efficiency of EVs Considered for Adoption in Humboldt County City Highway Vehicle Fuel Economy: Electric Fuel Economy: Electric Charge- Efficiency: Charge- Efficiency: Sustaining Charge-Depleting Sustaining Charge-Depleting mode (mpg) mode (Wh/mi) mode (mpg) mode (Wh/mi) Conventional 25.9 37.6 PHEV10 57.2 229.0 56.1 226.2 PHEV40 48.4 227.3 50.9 225.5 EREV30 49.3 213.0 51.3 215.5 EREV40 51.6 217.5 52.6 215.9 BEV100 190.7 206.5 Source: Santini et al. (2010)

Figure 28: Tank-to-Wheel Efficiency (BTU/mi) of EVs in Charge-Depleting Mode Versus a Conventional Vehicle

Average EV

BEV100

EREV40

EREV30

PHEV40

PHEV10

Convenonal

0 500 1,000 1,500 2,000 2,500 3,000 3,500 4,000 Convenonal PHEV10 PHEV40 EREV30 EREV40 BEV100 Average EV BTU/mi 3804 1014 1009 957 966 893 968

Source: SERC staff analysis

92 Figure 29: Fuel Cost ($/mi) of Electric Vehicles in Charge-Depleting Mode Versus a Conventional Vehicle

Average EV

BEV100

EREV40

EREV30

PHEV40

PHEV10

Convenonal

$0.00 $0.02 $0.04 $0.06 $0.08 $0.10 $0.12 $0.14 Convenonal PHEV10 PHEV40 EREV30 EREV40 BEV100 Average EV $/mi $0.122 $0.036 $0.04 $0.034 $0.034 $0.03 $0.034

Source: SERC staff analysis

To estimate GHG emissions a gasoline well-to-wheel emission factor of 97.8 g/MJ was assumed (CARB, 2009). For electricity generation, several emission factors estimated by the REPOP model were used (Table 18). These are based on different scenarios of electricity generation mixes in Humboldt County. Assuming an emission factor of 0.269 tCO2e/MWh (2010, Business-as-usual scenario) suggests that a conventional vehicle results in about five times the GHG emissions as the average electric vehicle does in charge-depleting mode (Figure 30). Emission factors for electricity generation are based on plant-to-wheel estimates, and not well-to-wheel.

93 Table 18: REPOP Estimated Retail Cost of Electricity and Power Plant Emissions Factor Based on Different Electricity Generation Scenarios in Humboldt County

Retail Generation Generation Year Generation Scenario Description Cost Emissions Scenario ($/MWh) (tCO2e/MWh) 2010 Business-as-usual • Natural Gas: 63.5% 122.4 0.269 • Biomass: 31.6% • Hydroelectric: 4.8% • Photovoltaic: 0.1% 2010 Medium renewables • Wind: 50 MW 123.5 0.152 penetration • Wave: 15 MW • Biomass: +20 MW • Hydroelectric: +10 MW • PV: +0.5 MW • EV penetration: 20% • Heat Pump penetration: 20% • Efficiency penetration: 20% 2010 High renewables • Wind: 125 MW 128.5 0.035 penetration • Wave: 30 MW • Biomass: +70 MW • Hydroelectric: +13 MW • PV: +1 MW • EV penetration: 38% • Heat Pump penetration: 38% • Efficiency penetration: 100% • Storage: 15 MW 2020 Business-as-usual Similar to 2010, Business-as-usual 121.6 0.275 2020 Medium renewables Similar to 2010, Medium renewables 122.7 0.162 penetration penetration 2020 High renewables Similar to 2010, High renewables 128.0 0.043 penetration penetration 2030 Business-as-usual Similar to 2010, Business-as-usual 120.9 0.281 2030 Medium renewables Similar to 2010, Medium renewables 123.3 0.178 penetration penetration 2030 High renewables Similar to 2010, High renewables 127.4 0.050 penetration penetration Source: SERC staff analysis

94 Figure 30: GHG Emissions (gCO2e/mi) of EVs in Charge-Depleting Mode Versus a Conventional Vehicle

Average EV

BEV100

EREV40

EREV30

PHEV40

PHEV10

Convenonal

0 50 100 150 200 250 300 350 400 Convenonal PHEV10 PHEV40 EREV30 EREV40 BEV100 Average EV gCO2e/mi 392 80 80 75 76 70 76

Source: SERC staff analysis

The Metropolitan Transportation Commission estimates that in 2008 annual vehicle miles traveled per capita in the North Coast region of California was about 10,500 miles. Based on California Department of Transportation and California Department of Motor Vehicles data from 1999 to 2007, average annual vehicle miles traveled per vehicle in Humboldt County was about 11,000 miles, or 30 miles per day, which was assumed for this analysis. National utility factors developed by EPRI were applied to determine the fraction of miles driven by plug-in hybrid electric vehicles that are derived from electricity (Figure 31). Utility factor is primarily dependent on annual vehicle miles traveled and all-electric range. In general, plug-in hybrid electric vehicles that are driven on many long trips, say beyond the 30-mile daily average in Humboldt County, will have lower utility factor. Other parameter assumptions that were used to estimate total energy use and costs use are included in Table 19 and Table 20.

95 Figure 31: Plug-In Hybrid Electric Vehicle Utility Factor as a Function of All-Electric Range and Annual Vehicle Miles Travelled, Assuming Nightly Charging

Source: EPRI (2007)

96 Table 19: Lifecycle Benefit / Cost Analysis Parameter Assumptions

Category Parameter Units Value Source Battery Battery Average Lifetime yr 10 Santini et al., 2011 Battery Efficiency (Round-trip) % 90 Santini et al., 2011 Battery Warranty mi 100000 Chevrolet; Nissan, 2011 Battery Warranty yr 8 Chevrolet; Nissan, 2011 Charging Charger Efficiency (Level 1) % 95 Santini et al., 2011 Charger Efficiency (Level 2) % 85 Santini et al., 2011 Charge Rate (Level 1) (110-120V, 15- 20A) kW 1.44 INL (DOE), 2008 Charge Rate (Level 2) (240V, 40A) kW 3.3 INL (DOE), 2008 Charge Rate (Level 3) (480V, 80A) kW 38.4 INL (DOE), 2008 Charger Capital Cost (Level 2) $ 2272 Santini et al., 2010 Vehicle Charges per day charge/day 1 Santini et al., 2011 Driving Days day/yr 360 Santini et al., 2011 Discount Rate Discount Rate % 6.4 Santini et al., 2011 Emissions Gasoline Well to Wheels CO2e g/MJ 97.800 CARB, 2009

Electricity Generation (BAS, 2010) tCO2e/MWh 0.26893 REPOP estimate, 2011

Elec. Gen. (BAS, 2020) tCO2e/MWh 0.27518 REPOP estimate, 2011

Elec. Gen. (BAS, 2030) tCO2e/MWh 0.28122 REPOP estimate, 2011

Elec. Gen. (MRP, 2010) tCO2e/MWh 0.15208 REPOP estimate, 2011

Elec. Gen. (MRP, 2020) tCO2e/MWh 0.16170 REPOP estimate, 2011

Elec. Gen. (MRP, 2030) tCO2e/MWh 0.17768 REPOP estimate, 2011

Elec. Gen. (HRP, 2010) tCO2e/MWh 0.03546 REPOP estimate, 2011

Elec. Gen. (HRP, 2020) tCO2e/MWh 0.04258 REPOP estimate, 2011

Elec. Gen. (HRP, 2030) tCO2e/MWh 0.05018 REPOP estimate, 2011 EVs PHEV10 Weight % 25 Santini et al., 2011 PHEV40 Weight % 13 Santini et al., 2011 EREV30 Weight % 13 Santini et al., 2011 EREV40 Weight % 25 Santini et al., 2011 BEV100 Weight % 25 Santini et al., 2011 Utility Factor (PHEV10, 10k VMT) % 16.67 EPRI, 2007 Utility Factor (PHEV30, 10k VMT) % 63.2 EPRI, 2007 Utility Factor (PHEV40, 10k VMT) % 71.11 EPRI, 2007 Parameter Definitions BAU: business-as-usual MRP: medium renewables penetration HRP: high renewables penetration VMT: vehicle miles traveled Source: SERC staff analysis

97 Table 20: Lifecycle Cost / Benefit Analysis Parameter Assumptions

Category Parameter Units Value Source EVs BEV Maintenance Cost $/mi 0.0077 EPRI, 2004 PHEV Maintenance Cost $/mi 0.02671 EPRI, 2004 CV Maintenance Cost $/mi 0.03598 EPRI, 2004 CV Maintenance Cost $/mi 0.0454 BTS, 2009 Tire Maintenance Cost $/mi 0.0083 BTS, 2009 Vehicle Lifetime year 17 CARB, 2009. Fuel Gasoline Price (2010-2030) $/gal 4.00 Humboldt Economic Index, 2011 Retail Electricity (BAU, 2010) $/kWh 0.12240 REPOP estimate, 2011 Retail Electricity (BAU, 2020) $/kWh 0.12164 REPOP estimate, 2011 Retail Electricity (BAU, 2030) $/kWh 0.12092 REPOP estimate, 2011 Retail Electricity (MRP, 2010) $/kWh 0.12347 REPOP estimate, 2011 Retail Electricity (MRP, 2020) $/kWh 0.12274 REPOP estimate, 2011 Retail Electricity (MRP, 2030) $/kWh 0.12329 REPOP estimate, 2011 Retail Electricity (HRP, 2010) $/kWh 0.12853 REPOP estimate, 2011 Retail Electricity (HRP, 2020) $/kWh 0.12796 REPOP estimate, 2011 Retail Electricity (HRP, 2030) $/kWh 0.12737 REPOP estimate, 2011

Humboldt County Driving Fraction (Freeway) % 60 Estimated from CADOT, 2009 Driving Fraction (City) % 40 Estimated from CADOT, 2009

Estimated from CADOT, 2009 and Average Annual Miles Driven mi/yr 11,000 CADMV, 2008

Estimated from CADOT, 2009 and Average Daily Driving Distance mi/day 30 CADMV, 2008 City Driving Schedule UDDS Santini et al., 2011 Highway Driving Schedule HWFET Santini et al., 2011 Estimated from CADMV, 2001- New Vehicle Sales (CA) % 7.89 2007 Estimated from CADMV, 2001- Fraction light duty trucks % 56.7 2007 Estimated from CADOT, 2009 and Average Fuel Efficiency of CV mi/gal 18 CADMV, 2008 Average Fuel Eff. of New CV (City) mi/gal 25.9 BTS, 2009 Average Fuel Eff. of New CV (Hwy) mi/gal 37.6 BTS, 2009 Incentives Federal Tax Credit (BEV) $ 7500 U.S. Department of Energy, 2010 Federal Tax Credit (PHEV) $ 7500 U.S. Department of Energy, 2010 Parameter Definitions BAU: business-as-usual MRP: medium renewables penetration HRP: high renewables penetration CV: conventional vehicle Source: SERC staff analysis

98 Table 21 shows estimated annual electricity and gasoline consumption, fuel costs, and emissions based on an annual vehicle miles traveled of 11,000 miles, a gasoline price of $4 per gallon, an electricity price of $0.12 per kWh, and emission factors corresponding to a 2010 business-as-usual generation mix. On average an EV reduces annual gasoline consumption by 256 gallons (78 percent), annual fuel cost by $792 (60 percent), and net emissions by 2.78 tCO2e (66 percent). Combining the projected penetration rates previously discussed for Humboldt County with average electric vehicle energy demand leads to the county level estimates in Table 22. It is assumed that each electric vehicle adopted in Humboldt County is purchased as a new vehicle and displaces the purchase of a new mid-sized conventional vehicle (for example a Toyota Camry), with its respective fuel economy. An electric vehicle is assumed to be retired from the county vehicle fleet after a lifetime of 17 years.

An additional 15 MWh and 30 MWh of daily electricity would be demanded from the grid under Medium and High penetration scenarios, respectively, in 2020. This would increase substantially to 126 MWh and 252 MWh of daily electricity consumption in 2030. To simulate vehicle charging, Level 1 charging (1.44 kW charge rate) and power demand profiles similar to those developed in an NREL study (Figure 32), were assumed. The four EV charging profiles are intended to represent boundary cases and probable charging scenarios. These were scaled to fit projected EV energy demand in Table 22.

Table 21: Estimated Vehicle Electricity and Gasoline Use and Associated Cost, and CO2e Emissions Electricity Use Gasoline Use

Vehicle kWh/day kWh/year $/year tCO2e/year gal/day gal/year $/year tCO2e/year Conventional 0.00 0 0 0.00 0.91 328 $1,313 4.24 PHEV10 1.49 535 $65 0.15 0.44 159 $637 2.06 PHEV40 6.31 2271 $276 0.62 0.17 62 $250 0.81 EREV30 5.32 1914 $233 0.53 0.22 79 $315 1.02 EREV40 6.04 2174 $264 0.60 0.17 60 $239 0.77 BEV100 7.85 2826 $344 0.78 0.00 0 $0 0.00 Average EV 5.30 1907 $232 0.52 0.20 72 $290 0.93

Source: SERC staff analysis

99 Table 22: Estimated Cumulative Electric Vehicles in Humboldt County in 2020 and 2030 and Associated Daily Electricity and Gasoline Demand, Under Medium, High, and Max Penetration Scenarios Projected Cumulative EVs Projected Daily EV Projected Daily EV Gas in Humboldt County Energy Consumption Consumption (gal) (MWh) Scenario Medium High Max Medium High Max Medium High Max 2020 2,833 5,666 10,731 15 30 57 570 1,139 2,158 2030 23,758 47,516 65,670 126 252 348 4,778 9,555 13,206

Source: SERC staff analysis

Figure 32: Four Vehicle Charging Scenarios Developed by NREL to Evaluate PHEV-Charging Impacts on the Utility System Operations Within the Xcel Energy Colorado Service Territory

EV Charing 130 Scenarios 120 110 100 90 80 70 60 50

kW/100 40 Vehicles 30 20 10 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Hour Uncontrolled Delayed Off-Peak Continuous

Source: NREL (2007)

100 A.2 References

Bureau of Transportation Statistics (BTS; U.S. Department of Transportation Research and Innovative Technology Administration) (2010). National Transportation Statistics, Table 4-23: Average Fuel Efficiency of U.S. Passenger Cars and Light Trucks. 2009. Table 3-14: Average Cost of Owning and Operating an Automobile.

California Air Resources Board (2009) Detailed California-Modified GREET Pathway for California Reformulated Gasoline (CaRFG). Release Date: January 12, 2009. http://www.arb.ca.gov/fuels/lcfs/011209lcfs_carfg.pdf [Accessed September 2010].

California Department of Transportation, Division of Transportation System Information (2009). 2008 California Motor Vehicle Stock, Travel and Fuel Forecast.

California Department of Transportation, Division of Transportation System Information (2002-2010). 2001-2009 California Public Road Data.

Chevrolet. Chevrolet Volt Warranty Information. http://www.chevrolet.com/volt/ [Accessed May 10, 2011].

Electric Power Research Institute (EPRI), Natural Resource Defense Council (NRDC), 2007. Environmental Assessment of Plug-In Hybrid Electric Vehicles, Volume 1: Nationwide Greenhouse Gas Emissions. http://mydocs.epri.com/docs/CorporateDocuments/SectorPages/Portfolio/PDM/P HEV-ExecSum-vol1.pdf. [Accessed December 7, 2010].

Electric Transportation Engineering Corporation (ETEC), 2010. Long-Range EV Charging Infrastructure Plan For Western Oregon (Draft, Version 1.3).

Humboldt Economic Index (Humboldt State University). 2011 Monthly Gasoline Prices for Eureka. http://www.humboldt.edu/econindex/ [Accessed May 10, 2011].

Morrow, K., et al (Idaho National Laboratory, U.S. Department of Energy). 2008. U.S. Department of Energy Vehicle Technologies Program – Advanced Vehicle Testing Activity. Plug-in Hybrid Electric Vehicle Charging Infrastructure Review.

National Renewable Energy Laboratory (2007) Parks, K., P. Denholm, and T. Markel. Costs and Emissions Associated with Plug-in Hybrid Electric Vehicle Charging in the Xcel Energy Colorado Service Territory. Technical Report NREL/TP-640-41410

Nissan USA. Nissan LEAF Warranty Information. http://www.nissanusa.com/leaf-electric- car/warranty-info/index. [Accessed May 10, 2011].

Santini, D., A. Vyas, D. Saucedo, and B. Jungers (2010) (Argonne National Laboratory and Electric Power Research Institute). 2010. Where Are the Market Niches for Electric Drive Vehicles?

U.S. Department of Energy. U.S. Government Source for Fuel Economy Information. http://www.fueleconomy.gov/feg/taxphevb.shtml [Accessed May 10, 2011].

101

APPENDIX B: Alternative Transportation Fuel Pathways

B.1 Hydrogen Transit Assessment

There are several hydrogen transit demonstration projects in the US and throughout the world. The National Renewable Energy Laboratory (NREL) has evaluated fuel cell bus demonstration projects funded by the Department of Transportation’s Federal Transit Administration (FTA) since 2000. NREL evaluates the buses, hydrogen infrastructure and the transit authority’s experience in implementing the program. One of the projects currently under evaluation is the Zero Emission Bay Area demonstration, which will include twelve next-generation Van Hool chassis buses with UTC Power fuel cells and hybrid propulsion systems. This demonstration will be led by AC Transit and will also include other transit agencies in the area such as Santa Clara VTA, Sam Trans, Golden Gate Transit, and San Francisco Municipal Transportation Agency. The project includes the construction of a new fueling station. Some of the hydrogen will be generated onsite using renewable electricity, and the balance of the hydrogen will be trucked in as a liquid. This project is currently the largest in the US (Eudy, Chandler & Gikakis, 2010). The FTA is also overseeing the National Fuel Cell Bus Program, which will provide $49 million in cost share for developing and demonstrating new transit buses. The funds will be provided for eleven buses. The program will expand the number of fuel cell power system manufacturers under evaluation and will demonstrate buses of different sizes and different hybrid propulsion designs. It will ideally be a final step before the introduction of commercial fuel cell buses into transit in the US. The buses will be located mostly on the East Coast, with four going to Connecticut, three to New York, and one each to Massachusetts and South Carolina. The final two will be in California, one each in San Francisco and Palm Desert (Eudy, et al., 2010). The largest fleet of fuel cell buses in the world is located in British Columbia, Canada. The fleet of 20 buses was introduced for the 2010 Winter Olympic Games in Whistler. The buses are from New Flyer and use Ballard fuel cells; the hybrid propulsion systems are from ISE/Siemens. The fueling station built to serve this fleet is the largest of any transit system. Transit systems and vehicle fleets are an ideal place for the introduction to hydrogen fuel cell vehicles because of the central fueling and maintenance infrastructure. They also expose a large number of people to the technology that would otherwise not have access to it and offer many opportunities for public outreach. Switching to fuel cell buses also has positive impacts on local air quality. The only emission from a fuel cell bus is water vapor, completely eliminating the particulates and smells associated with diesel buses. Fuel cell buses are also much quieter than conventional diesel buses.

102 B.1.1 Hydrogen Fuel Cell Bus Analysis Methodology

Mass transit makes up a small percentage of Humboldt County’s overall transportation energy consumption, accounting for only 1.5 percent of diesel fuel consumed in the county annually. Humboldt County uses approximately 195,000 gallons of diesel a year to fuel its three transit systems. Redwood Transit System is the largest system in the county. It operates Monday thru Saturday from Scotia in the south to Trinidad in the north, with weekday service to Willow Creek. Within the south county it operates from Garberville to Myers Flat. Eureka Transit System serves the city of Eureka and the Arcata and Mad River Transit System serves the city of Arcata. Annually the three systems travel approximately 1,250,000 miles and have a ridership of over one million passengers. Greenhouse gas emissions associated with diesel fuel use are approximately 433,000 pounds of CO2 equivalent annually. The purpose of this analysis was to investigate the cost of a hydrogen-based transit system in Humboldt County and the greenhouse gas emission impacts. Currently there are several ways to generate hydrogen. The most common method is steam reformation of methane (the main component of natural gas). Hydrogen can also be generated from water via electrolysis. Electrolysis uses electricity to split water into hydrogen and oxygen. If the electricity used for electrolysis is renewable, the hydrogen fuel can be considered a renewable fuel. Generating hydrogen with renewable energy is a way of storing renewable energy and is particularly useful if surplus renewable energy is available. For this analysis only hydrogen generated via electrolysis was considered, as generating hydrogen from fossil fuels does not satisfy the goals of the RESCO project. This analysis used a Microsoft Excel® model to generate the desired outputs. The user can define the following inputs to the model: • Total transit miles travelled annually in Humboldt County (1,250,000) • Fuel economy for the fuel cell buses (7.77 miles/kg) • Fuel economy for diesel buses (6.41 miles/gallon) • Price of diesel ($3.5/gallon) • Price of electricity ($0.124/kWh) • Carbon intensity of electricity (0.304lbs/kWh) • Energy intensity to generate and compress hydrogen (60 kWh/kg) • Operations and maintenance costs for diesel buses ($4.03/mile includes labor costs) • Operations and maintenance costs for fuel cell buses ($4.43/mile includes labor costs)

Most of the variables were held constant for the analysis, though several were changed for a sensitivity analysis as discussed below. The number of miles traveled by the fuel cell buses and the fuel efficiency of the buses was used to determine the hydrogen generation requirements. The fueling station was

103 sized to meet the maximum daily hydrogen requirements plus an additional 20 percent. Various levels of hydrogen bus penetration were considered, from a small hydrogen system covering 10 percent of the total miles traveled to full-scale hydrogen system covering 100 percent of the miles traveled. Costs associated with a hydrogen transit system include: upfront costs for building a fueling station and purchasing buses, ongoing operations and maintenance costs for the fueling station and hydrogen buses, and electricity costs associated with generating and compressing hydrogen. Costs associated with building, operating and maintaining the fueling station came from a California Hydrogen Highway economic report (Economy Topic Team, 2005). The costs for operations and maintenance of the fuel cell buses were based on reports by AC Transit for their fuel cell bus pilot program and were adjusted down to account for the learning curve that will likely cause costs to decrease as the technology matures and becomes more reliable. The cost used for the fuel cell buses, $750,000, is a number estimated by AC Transit to be the cost ceiling necessary for commercialization of fuel cell buses. Current fuel cell bus costs are closer to $2 million. The cost of electricity used in the analysis is based on a likely 2030 scenario as determined using the REPOP model. Costs for the current diesel bus system were based on data provided by Humboldt Transit Authority and A&MRTS. The cost of diesel fuel was assumed to be the same as the current cost. A sensitivity analysis on fuel costs is included to address how this assumption affects the final results. The amount of carbon dioxide offset by the hydrogen system is the amount of carbon dioxide that would have been produced by the offset diesel fuel, minus the amount of carbon dioxide produced while generating electricity to produce hydrogen. The cost for the offset carbon dioxide is the difference in cost between BAU and the hydrogen scenario.

B.2 Forest-Based Biofuels Assessment

An assessment was conducted to determine which forest-based biofuel pathways are closest to commercialization. It was determined that the Fischer-Tropsch biodiesel pathway and the cellulosic ethanol pathway are most likely to be successful. Each of these pathways is described below. B.2.1 Fischer-Tropsch Biodiesel

For lignocellulosic feedstocks, gasification is generally considered to be the most convenient conversion method because of its applicability to both electricity and liquid fuel production. Gasification is a general term used to describe gaseous fuel production by partial oxidation of a solid fuel (NTNU, 2005). The product of biomass gasification is synthesis gas (syngas), consisting primarily of hydrogen (H2), carbon monoxide (CO), carbon dioxide (CO2) and methane (CH4). The thermochemical conversion is carried out under high temperatures (1,500°F to 2,000°F) with a controlled amount of oxygen and/or steam. Syngas can serve as a feedstock for catalytic synthesis into liquid fuels (for

104 example, ethanol, gasoline, diesel) or be used as a substitute for natural gas in cogeneration engines, turbines and boilers to produce electricity and/or heat. Gas-to-liquids is a refinery process used to convert gaseous hydrocarbons into longer- chain hydrocarbons, such as automotive fuel. This can be accomplished by either bacterial fermentation or, more commonly, catalytic synthesis. Catalytic synthesis is most often carried out through the Fischer-Tropsch process, which is capable of synthesizing a number of automotive fuels. Though ethanol production is possible through Fischer-Tropsch, the process is usually applied to diesel fuel production. The Fischer-Tropsch process is widely considered to be the most promising technology available today for woody biomass gas-to-liquids fuel production (NTNU, 2005).

Fischer-Tropsch synthesis is a carbon chain building process whereby CH2 groups are attached to the ends of carbon chains. The exact sequence and behavior of the chemical reactions that take place have been subject to controversy since the 1930’s (NTNU, 2005), but can generally be described by the overall equation: ! !"# + ! + ! à ! ! + !! ! ! ! ! ! ! The reactions are highly exothermic and require sufficient cooling to maintain stable reaction conditions. Overheating results in undesirably high yields of methane production and an increased rate of catalyst deactivation due to sintering and fouling (Dry, 2002). A number of catalysts can be used, but most commonly transition metals such as cobalt, iron or ruthenium are used. A wide range of products result from the Fischer-Tropsch reaction, including olefins, paraffins, and oxygenated products (alcohols, aldehydes, acids and ketones). Fischer- Tropsch synthesis invariably produces a small portion of methane, which is usually reformed with steam into H2 and CO, and fed back into the reactor. The fuel produced directly out of the Fischer-Tropsch reactor is termed straight-run to distinguish itself from the fuel produced after synthesis from upgraded hydrocarbons. The output of desired products can be greatly increased by selecting an appropriate reactor type, catalyst, pressure range and temperature mode. Fischer-Tropsch reactions are classified into two modes: low and high temperature which take place at temperature ranges of 200-240°C (392-482°F) and 300-350°C (572-662°F), respectively (Dry, 2002).

Fischer-Tropsch diesel synthesis works best with a high capacity low temperature slurry bed reactor and a cobalt or iron catalyst. Along with producing about 20 percent straight-run diesel, this reaction produces excess paraffin wax, which is usually hydrocracked and fed back into the reactor until extinction. The straight-run diesel has a cetane number of 75 to 80 whereas diesel hydrocracked from wax has a cetane number of 70 to 75 (Leckel, 2009). The resulting mixture of straight-run and hydrocracked diesel is a fuel with cetane number of about 70. Compared to conventional petro-diesel, which is required to have cetane of 45 or higher, Fischer-Tropsch diesel is of higher quality, lower aromatic content and free of sulfur. This allows Fischer-Tropsch diesel to be used as a blending stock to upgrade lower quality fuels (Dry, 2002).

105 B.2.2 Cellulosic Ethanol

The biochemical approach is most often utilized for bioethanol production through hydrolysis and subsequent fermentation. While being slower than a chemical reaction, the biological process has several advantages, including higher specificity, higher yields, lower energy costs and greater resistance to catalytic poisoning (NTNU, 2005). Ethanol must be blended with other fuels when used as a transportation fuel. It has a high octane number and is commonly used as a petro-fuel additive to improve engine performance and emissions (NTNU, 2005). Most gasoline is blended with five to ten percent ethanol, but with minor engine modifications (for example, Flex Fuel Vehicles), blends of up to 85 percent can be utilized.

The fermentation of glucose (C6 sugar) is a well-known process and has been practiced for at least 6,000 years, beginning with early beer-making societies such as the Sumerians, Egyptians and Babylonians (NTNU, 2005). Most of the ethanol produced today (approximately 90 percent), as was the case in ancient times, comes from sugar and starch crops. These crops have high food value and low sugar yield per acre compared to cellulose and hemi-cellulose contained in lignocellulosic biomass, which is the most prevalent form of sugar in nature (NTNU, 2005). Unlike sugar and starch crops, which contain C6 sugars that are 100 percent convertible by normal yeasts, lignocellulosic biomass contains the C5 sugars xylose and arabinose as well as the C6 sugars glucose, galactose and mannose. It is estimated that 40 percent of sugars contained in lignocellulosic biomass cannot be metabolized by yeast. This 40 percent disadvantage largely contributes to why lignocellulosic feedstocks remain far from being competitive with starch and sugar crops for ethanol production (NTNU, 2005).

A pretreatment step prior to hydrolysis is required to remove lignin content and hydrolyze hemicellulose. The cellulose remaining after pretreatment can then be hydrolyzed into fermentable sugars (that is, glucose). Hydrolysis is a chemical reaction used to break down polymers, during which water molecules are split into hydrogen cations and hydroxide anions. In the case of cellulose hydrolysis, the following reaction takes place:

!!!!"!! ! + !!!! à !!!!!"!! This reaction is traditionally catalyzed by sulfuric acid (either concentrated or dilute), but recent developments favor enzymatic hydrolysis using cellulase. Cellulase is a class of enzymes that catalyze cellulose hydrolysis and offers many advantages over acid hydrolysis, including lower maintenance costs, high yields and no corrosion issues.

Projected costs for enzymatic hydrolysis are estimated to be four times lower than concentrated acid hydrolysis, and three times lower than dilute acid hydrolysis (TSS Consultants, 2010). These developments, along with lower maintenance costs, no corrosion, and a lower environmental impact contribute to the belief among many experts that enzymatic hydrolysis is the most cost-effective method for producing ethanol in the long term (NTNU, 2005). The production cost of ethanol utilizing

106 enzymatic hydrolysis is currently estimated at $2.24 per gallon for a 2,200 Bone Dry Ton (BDT) per day plant (NTNU, 2005).

A variety of microorganisms are capable of fermenting carbohydrates into ethanol under oxygen-free conditions, including bacteria, yeasts, and fungi. The reactions that take place are the following (Hamelinck, 2004):

3C5H10O5 à 5C2H5OH + 5CO2

C6H12O6 à 2C2H5OH + 2CO2

Fermentation of hexose sugars in lignocellulosic crops only became possible during the late 1800’s. Pentose sugars, which represent the majority of available sugar in lignocellulosic crops, were not capable of being fermented until the 1980’s, when a number of wild yeasts were discovered that were able to convert xylose to ethanol (Hamelinck, 2004). While bacteria are capable of faster fermentation than yeasts (minutes compared to hours), all microorganisms have their limitations, including at least one of the following: inability to process both pentoses and hexoses, low ethanol yields, co-production of cell mass at the expense of ethanol, and/or slow extermination of the microorganism population by oxygen free conditions (Hamelinck, 2004). Current fermentation configurations include separate reactors for glucose and xylose fermentations, since different microorganisms are needed to process the different sugars. Genetic engineering is expected to develop bacteria and yeasts that are capable of fermenting both xylose and glucose, and could have the potential to utilize the remaining sugars (mannose, galactose, arabinose) (Hamelinck, 2004).

B.2.3 Parameter Values and Results

This section contains inputs and results related to the forest-based biofuels assessment discussed in Section 2.3.2.3. Table 23 contains customized, non-default parameters used in the Greenhouse Gases, Regulated Emissions, and Energy Use in Transportation (GREET) Model. Original equipment manufacturer plug-in hybrid electric vehicle and battery-electric vehicle plans between 2010 and 2012 are listed in Table 24. According to original equipment manufacturer plans, 8 plug-in hybrid electric vehicle models and 17 battery-electric vehicle models are planned for market introduction by 2012.

Table 25 contains well-to-wheel energy and emission changes for several alternative fuel pathways relative to a baseline of conventional vehicles fueled with conventional gasoline. Metrics included in this comparison are fossil fuels, GHGs, and criteria pollutants. Battery-electric vehicles have the largest impact, reducing fossil fuel use and GHGs by 95 percent and 92 percent, respectively, followed by the Fischer-Tropsch Diesel (FTD) pathway, which results in reductions of 92 percent and 91 percent.

Similarly, Table 26 reports well-to-pump energy and emission changes for the same fuel pathways relative to the same baseline. On a well-to-pump basis the Fischer-Tropsch Diesel and Ethanol (via fermentation) pathways have the largest GHG impact, reducing emissions by 447 percent and 357 percent, respectively.

107 Table 23: GREET Fuel Pathway Parameters Customized to Humboldt County. Category Parameter Value Source Simulation Target year for simulation 2020 Maximum allowed Simulation vehicle types Passenger Cars Fischer- Tropsch Forest Residue: Diesel (FTD) Classification of biomass for FTD 100% Mass-based share of coal and biomass co-feeding to FTD Biomass: 100%, production Coal: 0% Forest Residue: Technologies for biomass-based Gasification and Ethanol ethanol plant Fermentation Ethanol Yield (gallons per dry Fermentation: 59 TSS Consultants, ton) gal 2010 Electricity co-produced for export Fermentation: - TSS Consultants, (kWh per gallon of ethanol) 3.475 2010 Electricity Mix for transportation use (user Generation defined) Biomass: 100% Natural Gas: 63.5%, Biomass: Estimated Mix for stationary use (user 31.6%, Others: generation mix for defined) 4.9% Humboldt County Hydroelectric: Estimated Shares of technologies for other 97.5%, generation mix for power plants (stationary use) Solar PV: 2.5% Humboldt County Share of feedstock for biomass Forest Residue: power plants 100%

Source: SERC staff analysis

108

Table 24: Automotive Manufacturer Plans for Plug-In Hybrid Electric Vehicles (PHEVs) and Battery Electric Vehicles (BEVs) Make Model All Electric Battery U.S. Target Range (mi) Size (kWh) Intro. Date

Plug In Hybrid Electric Vehicles

Audi A1 Sportback 31-62 2011 BYD Auto F3DM 60 2010 Fisker Karma 50 2010 Ford Escape 40 10 2012 General Motors Chevrolet Volt 40 16 2010 Hyundai Blue-Will 38 2012 Toyota Prius Plug-in 12.4-18.6 2012 Volvo V70 31 2012

Battery Electric Vehicles

BMW ActiveE 100 2011 BYD Auto e6 205 2010 Chrysler/Fiat Fiat 500 100 2012 Coda Automotive Coda Sedan 90-120 2010 Daimler Smart ED 72-90 2012 Mercedes Benz 120 35 2010 low volume Ford Focus 100 2011 Transit Connect 100 2010 Tourneo Connect 100 21 2011 Hyundai i10 Electric 100 16 2012 Mitsubishi iMiEV 100 16 2010 Nissan LEAF 100 24 2010 Rolls Royce Electric Phantom 2010 SAIC Roewe 750 125 2012 Tesla Motors Roadster 220 56 For sale now Model S 160, 230, 2011 300 Th!nk City 113 2010

Source: ETEC (2010)

109 Table 25: Wells-to-Wheels Energy (BTU/mi) and Emission (g/mi) Changes for Several Fuel Pathways

Biomass Biomass EtOH FFV: E85, E85, FFV: EtOH E85, FFV: EtOH (BTU/mior g/m) BEV: Electricity, Electricity, BEV: SI PHEV: CG and and CG PHEV: SI Electricity, Biomass Electricity, CIDI Vehicle: FT100, FT100, Vehicle: CIDI Baseline: CG and RFG and CG Baseline: Biomass (gasification) Biomass

Metric (fermentation) Biomass Total Energy 5,279 24.2% 83.4% 39.5% -11.8% -15.1% Fossil Fuels 4,909 -63.5% -78.6% -91.9% -47.2% -95.4% Coal 28 -73.8% -73.9% -100.0% -49.2% -100.0% Natural Gas 537 -41.9% -211.7% -92.1% -47.3% -95.6% Petroleum 4,345 -66.1% -62.2% -91.8% -47.1% -95.3% CO2 (w/ C in VOC & CO) 387 -64.7% -75.4% -91.8% -47.6% -95.2% CH4 0.484 -60.0% -86.4% -91.7% -46.3% -92.3% N2O 0.020 26.3% 145.3% -34.0% 61.1% 137.1% GHGs 405 -63.2% -72.5% -91.0% -46.0% -91.7% VOC: Total 0.265 -17.6% 3.3% -72.4% -48.7% -89.4% CO: Total 3.540 -22.9% -18.3% -91.8% -35.0% -90.3% NOx: Total 0.253 14.3% 99.6% -31.5% 38.9% 100.0% PM10: Total 0.054 -3.0% 111.6% -36.2% 17.4% 43.4% PM2.5: Total 0.030 -2.2% 60.7% -39.8% 8.7% 23.1% SOx: Total 0.086 -54.9% -51.6% -65.5% -38.2% -74.9% VOC: Urban 0.158 -18.1% -18.2% -74.2% -53.4% -99.7% CO: Urban 2.178 -25.2% -25.5% -92.7% -39.2% -100.0% NOx: Urban 0.071 -34.3% -59.8% -34.0% -48.2% -98.3% PM10: Urban 0.023 -15.8% -20.0% -32.3% -19.8% -44.5% PM2.5: Urban 0.013 -17.4% -26.0% -42.1% -27.2% -62.8% SOx: Urban 0.022 -65.0% -62.4% -90.8% -47.6% -96.4%

( percent change is relative to a baseline of conventional vehicles fueled with conventional and reformulated gasoline)

Source: SERC staff analysis

110 Table 26: Well-to-Pump Energy (BTU/mmBTU) and Emission (g/mmBTU) Changes for Several Pathways

ty,

Biomass Biomass g/mmBTU) EtOH FFV: E85, E85, FFV: EtOH E85, FFV: EtOH BEV: Electrici BEV: (BTU/mmBTUor SI PHEV: CG and and CG PHEV: SI Electricity, Biomass Electricity, CIDI Vehicle: FT100, FT100, Vehicle: CIDI Baseline: CG and RFG and CG Baseline: Biomass (gasification) Biomass

Metric (fermentation) Biomass Total Energy 264,241 115.8% 399.2% 322.7% 178.9% 859.8% WTP Efficiency 79.1% 63.7% 43.1% 47.2% 57.6% 28.3% Fossil Fuels 217,596 -22.9% -104.9% -47.5% -8.6% -41.2% Coal 6,650 -73.8% -73.9% -100.0% -20.8% -100.0% Natural Gas 128,522 -41.9% -211.7% -90.5% -18.7% -89.7% Petroleum 82,424 10.6% 59.2% 23.9% 8.2% 39.2% CO2 (w/ C in VOC & CO) 15,895 -368.7% -430.8% -523.0% -1.7% -8.2% CH4 113.413 -61.3% -88.3% -90.5% -15.4% -74.0% N2O 1.889 66.3% 366.3% -82.8% 388.9% 1868.8% GHGs 19,293 -310.9% -357.2% -446.6% 7.7% 36.9% VOC: Total 27.358 -8.7% 39.8% -86.1% -4.0% -19.0% CO: Total 13.806 103.1% 386.9% -21.1% 387.0% 1859.4% NOx: Total 44.051 27.1% 144.5% -31.8% 167.7% 805.8% PM10: Total 6.077 -6.4% 237.4% -53.9% 132.9% 638.5% PM2.5: Total 3.617 -4.3% 120.2% -46.0% 113.1% 543.5% SOx: Total 19.216 -53.7% -50.2% -55.9% -2.5% -12.0% VOC: Urban 15.434 -9.1% -9.3% -93.3% -20.3% -97.3% CO: Urban 2.909 -60.0% -108.0% -85.3% -17.7% -85.2% NOx: Urban 6.697 -56.2% -120.8% -82.8% -17.8% -85.5% PM10: Urban 1.350 -65.6% -83.2% -89.2% -17.9% -85.9% PM2.5: Urban 0.810 -64.7% -96.7% -88.6% -17.9% -85.9% SOx: Urban 4.494 -63.4% -60.4% -87.0% -17.9% -86.0%

( percent change is relative to a baseline of conventional and reformulated gasoline)

Source: SERC staff analysis

111 B.3 References

Dry, Mark E. "The Fischer–Tropsch Process: 1950–2000." Catalysis Today 71 (2002): 227- 41.

Economy Topic Team Report (2005) California 2010 Hydrogen Highway Network. January 5, 2005

Eudy, L., Chandler, K., & Gikakis, C., (2010) Fuel Cell Buses in U.S. Transit Fleets: Current Status 2010. Technical Report NREL/TP-56000-49379.

Hamelinck, Carlo, Geertje Van Hooijdonk, and Andre Faaij. Ethanol from Lignocellulosic Biomass: Techno-economic Performance in Short-, Middle-, and Long-term. Report 22, December, 2004.

Leckel, Dieter. "Diesel Production from Fischer-Tropsch: The Past, the Present, and New Concepts." Energy & Fuels 23 (2009): 2342-358. March, 2009.

NTNU, Norwegian University of Science and Technology: Vessia, Øyvind. Biofuels from Lignocellulosic Material: Technology, Potential and Costs, In the Norwegian Context 2010. December, 2005.

TSS Consultants: Schuetzle, Dennis, Gregory Tamblyn, Fredrick Tornatore. Alcohol Fuels from Biomass – Assessment of Production Technologies. California Energy Commission. August, 2010.

112 APPENDIX C: Energy Efficiency Analysis

This appendix provides a detailed description of the methodology used to carry out the energy efficiency analysis in the REPOP model. Potential energy savings were based on results of a statewide efficiency study conducted by Itron (2008). The Itron estimates were adapted to apply to Humboldt County as described below.

C.1 Methodology

The Itron (2008) study quantifies yearly (2007-2026) energy savings potential and associated costs by investor owned utility (IOU) service territory, Energy Commission climate zone (CZ), consumption sector, building type, end-use category, fuel type (electricity or natural gas), and efficiency measure type. The sectors and end-use categories included in the analysis are provided in the main body of the report in Table 1. Each end-use category is comprised of a number of efficiency measures that provide energy savings above a common base technology. Associated building types are listed in Table 27 below. Table 27: Building Types by Sector Sector Residential Commercial

Colleges Grocery Stores Health Lodging Large Office Single Family Building Small Office Multi-Family Type Refrigerated Mobile Home Retail Restaurants Schools Warehouses Miscellaneous

Source: Schatz Energy Research Center staff.

For each of the individual measures, three categories of savings potential are quantified: technical, economic, and market potential (Figure 33). Technical potential is the highest level of savings and refers to the savings that would be realized if all applicable and feasible measures were actually installed. Economic potential is a subset of technical potential that limits the applicable and feasible measures to those that are cost-effective. Cost-effectiveness is determined using a total resource cost test, which is a ratio of avoided costs to incremental measure costs (neglecting programmatic costs). A measure with a total resource cost greater than one would be considered cost-effective. Market

113 potential is a further subset that takes into account factors such as customer awareness, customer willingness to adopt the measure, and market barriers. Figure 33: Categories of Savings Potential

Source: Itron (2008)

This study focused on market potential, as these are the measures that are most likely to be implemented. Because market potential is influenced by a number of uncertain factors, the Itron analysis considered 10 scenarios that reflect various incentive levels and total resource cost restrictions. A full description of these scenarios can be found in Itron (2008). Personal communications with Itron indicated that the Base, Mid, and Full total resource cost restricted scenarios reflect the most realistic estimates of achievable potential. These scenarios demonstrate a wide range of conservative market potential estimates that serve as the basis of the SERC analysis (Table 28).

114 Table 28: Total Resource Cost (TRC) Restricted Market Potential Scenarios Scenario Name Scenario Description Base Incentive TRC Includes measures incentivized in the 2004-2005 program cycle Restricted with incentive levels that were available in 2006. Restricted to measures with a TRC ≥ 0.85.

Mid Incentive TRC Includes all measures analyzed in the study with incentives Restricted halfway between those that were available in 2006 and full incremental costs. Restricted to measures with a TRC ≥ 0.85.

Full Incentive TRC Includes all measures analyzed with incentives set to full Restricted incremental costs. Restricted to measures with a TRC ≥ 0.85.

Source: Itron (2008)

For each measure under consideration, the Itron databases provide estimates for measure, program, and incentive costs borne by the IOU, as well as the associated energy savings (kWh or therms). The databases also provide estimates for the number of yearly installations of each measure from 2007-2026. The installations are further classified as device retrofit, equipment conversion, or replace on burnout. The distinction between conversion and retrofit is that a conversion is replacing existing equipment with an alternate technology (for example replacing a furnace with a heat pump) while a retrofit is upgrading existing equipment with a more efficient device of the same type (that is upgrading and HVAC system from SEER 13 to SEER 15). Replace on burnout means the energy efficient equipment is installed when the existing equipment fails. These decision types have various cost implications and are therefore modeled differently in the Itron analysis.

The Itron analysis was performed for each California Energy Commission (Energy Commission) climate zone because some measures, such as heating, ventilation, and air conditioning (HVAC), are largely weather dependent. In order to approximate energy savings potential for weather dependent measures in Humboldt County, this analysis focuses on the PG&E service territory in Energy Commission climate zone 1. Geographically, Humboldt County is split approximately evenly between zones 1 and 2 (Figure 34). Approximately 94 percent of the county’s population, however, lies in CZ 1 so energy savings potential for this zone, scaled by CZ 1 and Humboldt County’s population, is used in this analysis. In most cases, Itron’s assumptions are followed regarding the installation of measures appropriate for CZ 1. Based on local knowledge of energy end-uses in Humboldt County, however, measures involving residential air conditioning and pool pumps are excluded from this study.

For weather insensitive measures, Itron reported costs and energy savings estimates for the entire PG&E service territory. Population could not be used to scale these results because of the difficulty in estimating the population of an IOU service territory. Instead, residential and commercial building electricity and natural gas consumption

115 data were obtained from the Energy Commission and were used to scale costs and savings.

In the REPOP Model, the base incentive program described in Table 28 is assumed to continue into the future regardless of any action taken by the local community. Therefore, a program level of zero percent corresponds to this base scenario. The full incentive program is assumed to be upper limit of efficiency efforts, so a program level of 100 percent is used to represent this scenario. Energy savings and costs associated with Humboldt County’s efficiency program are assumed to scale linearly between zero and 100 percent. Figure 34: Humboldt County Energy Commission Climate Zones

Source: California Energy Commission, http://www.energy.ca.gov/maps/building_climate_zones.html

C.2 References Itron. 2008. California Energy Efficiency Potential Study. CALMAC Study ID: PGE0264.01.

116 APPENDIX D: Regional Energy Sector Optimization Model

This appendix provides detailed descriptions of Regional Energy Planning Optimization (REPOP) Model. In section D.1, each supply submodel is detailed, demonstrating how the hourly availability of electricity production for that resource is modeled. Section D.2 covers the balance support submodels, which are the technologies associated with accommodating hour-to-hour discrepancies between demand and available supply. The demand submodels are described in section D.3. In section D.4, the energy balance algorithm is explained, which is the methodology used to dispatch supply to meet demand in each hour of the year. Section D.5 describes the algorithm used in the optimization of the energy balance model. Finally, sections D.6 and D.7 contain tables of model assumptions and references, respectively.

D.1 Supply Submodels

D.1.1 Natural Gas

The natural gas submodel assumes that the PG&E Humboldt Bay Power Plant can produce power any hour of the year. This assumption is based on the modular nature of the plant, which consists of ten 16 MW internal combustion engines. Maintenance and forced outages need not occur on all of the engines at once. The scheduling of maintenance is assumed to occur during periods of low demand and/or high availability of power from other sources. Table 42 presents the technical and economic assumptions associated with this technology.

D.1.2 Wind

The wind submodel is based on three years of modeled hourly wind data from NREL’s Western Wind Dataset. The data were produced by 3Tier using mesoscale atmospheric modeling techniques covering the years 2004-2006. Data were available for several locations in the northern Cape Mendocino area, which represents the region of Humboldt County with the highest on-shore wind resource and closest proximity to existing transmission infrastructure (Figure 5 and Figure 6). Of the various data products provided by the dataset, the SCORE-lite corrected data are considered the most relevant to requirements of this analysis. These data represent a simulation of the output of 10 turbines instead of one. In addition, the SCORE-lite correction accounts for statistical deviations between the rated capacity of turbines and the output typically observed in practice (3Tier, 2010).

Model application:

• Given a source year (2004-2006) for the wind submodel and a plant capacity:

117 o The SCORE-lite corrected data are scaled to the capacity of the plant and the resulting time series represents the hourly availability of wind power to the Humboldt grid.

Table 43 presents the technical and economic assumptions associated with this technology.

D.1.3 Wave

Availability of power from a wave energy power plant is based on spectral wave data from the National Oceanic and Atmospheric Agency’s National Data Buoy Center and a Pelamis power matrix. Buoy 46022 is located approximately 14 nautical miles offshore from the mouth of the Eel River in water approximately 600 meters deep. A wave farm in the Humboldt County region would likely be located approximately 10 nautical miles from shore (PG&E, 2010a). The depth at this distance is approximately 140 meters, which puts the farm in sufficiently deep water that all wavelengths which can be captured by a wave energy converter are considered deep water waves (waves which have a wavelength less than twice the water depth). Hence, the observations at Buoy 46022 can reliably be used as a proxy for the wave environment at a distance of approximately 10 miles from shore. The buoy records spectral wave measurements at 10-minute intervals. A complete year of data are available for the following years: 1997, 1999, 2001-2003, 2005-2008.

Pelamis Wave Power Ltd. is a wave energy converter manufacturer based in Scotland and was the first company to deploy full-scale converters. They are also the only company that makes power conversion data from a deep-water device available to the public. These data, known as a power matrix, specify the amount of power produced by a device at a given sea state (Figure 35). The sea state is specified by the combination of the significant wave height (Hs) and energy period (Te). The formulae for calculating Hs and Te from spectral wave data follow.

!! = 4 !!

!!! !! = !!

! !! = (! )! !" !

Where Hs is significant wave height in meters, Te is energy period in seconds, !! is the n- th moment, ! is frequency in Hz, and ! is the spectral density in m2/Hz.

118 Figure 35: Power (kW) at Various Sea States Produced by the Pelamis Linear Attenuator

Source: SERC Staff based on Pelamis wave converter data (2010)

Model application:

• Given a source year (1997, 1999, 2001-2003, 2005-2008) for the wave submodel and a plant capacity:

o Calculate Hs and Te for every 10-minute interval of the year. o Use the Pelamis power matrix to estimate power production from a single WEC for each interval. o Average the data to an hourly interval, then scale to the plant capacity.

Table 44 presents the technical and economic assumptions associated with this technology. D.1.4 Biomass

Availability of power from biomass power plants are modeled based on daily production data from the DG Fairhaven Power Plant (16 MW) for the year 2009 (Marino, 2010). These data include gross and net electricity production, biomass and natural gas

119 fuel consumption, and explanatory data concerning scheduled and forced outages. The entire time series is plotted in Figure 36. The summer months in which the plant operated at a reduced level is directly related to a contract dispute between the plant and its power purchaser. These data are therefore ignored in the development of the biomass production model as they represent atypical circumstances for plant operation. Figure 36: Fairhaven Plant Production and Natural Gas Usage Data (Marino, 2010)

Fairhaven 2009 Average Daily Power Sold (MW) 12000 15 9000 10 Power Sold Forced Outage

Scheduled Outage 6000 Gas usage Gas Usage (therms) Average Daily Power(MW) 5 3000 0 0

Jan Mar May Jul Sep Nov Jan

Date

Source: SERC Staff (2010)

A composite stochastic time series model was developed to synthesize production data and outage occurrences from a biomass power plant. The Fairhaven data set was used to estimate model parameters. The authors recognize the potential hazard of using such a relatively small data set in stochastic model development. However, the model satisfied the following relatively simple, achievable criteria for model selection:

• The model should approximate observed daily variation in production unrelated to scheduled or forced outages. • The model should approximate scheduled and forced outages at the same frequency and magnitude as the observed outages.

D.1.4.1 Model Formulation The Fairhaven production data are converted into a daily capacity factor based on the maximum observed daily production in 2009. Data whose capacity factors are greater

120 than or equal to 80 percent are used to model the operation of the plant under normal operating circumstances. The data exhibited an unexplained, gradual trend over the course of the year. The cause of this trend was ignored, as the primary intent of the model is to reproduce the daily variation in plant output. So the data were de-trended and the residuals were used in the subsequent analysis.

The data exhibit clear autoregressive behavior as demonstrated by the autocorrelation function for the series (Figure 37). A first order autoregressive moving average model (p=1, q=1) was determined to adequately reproduce the autocorrelation structure of the series. This model was used to generate a synthetic series of biomass production under normal operating conditions (Figure 38). It is clear from the juxtaposition of the observed and synthetic time series inFigure 38 that the synthetic series has a larger variance than the observed. However, the mean and autocorrelation structure are the same, satisfying the model selection criteria. The model for normal operating conditions follows the following form: !!"#$ = !"#! 1,1 + !!"#$ where !!"#$ is the synthetic series, !"#!(1,1) is the time series produced by the fitted autoregressive moving average model, and !!"#$ is the mean value of the original normal operating time series before being de-trended.

Figure 37: Autocorrelation and Partial Autocorrelation Function for Residuals of a Detrended Time Series of Normal Operation (Capacity Factor >= 80 Percent) at Fairhaven Power

Autocorrelation Structure of Normal Operation Residuals 0.8 0.4 ACF 0.0

5 10 15 20 25 LAG 0.8 0.4 PACF 0.0

5 10 15 20 25 LAG

Source: SERC Staff (2010)

121 Data whose capacity factors are <80 percent (with the exception of scheduled maintenance) are considered low operating conditions and represent events when the plant is partially or fully curtailed. These conditions can arise for a multitude of reasons. For the purposes of this model, only the magnitude and frequency of the occurrences need to be reproduced by the model. To model the frequency of occurrence, the set of inter-arrival (for example days between) of low operating events are treated as a sample (Figure 39). A truncated gamma distribution is used to model the inter-arrival times, with the following form: ! = !"##" !, ! 150, ! ≥ 150 ! = !"#$%!!"" !"#$% ! , ! < 150 where !, ! are the parameters of the gamma distribution, ! is the result of a random draw from that distribution, and !!"#$%!!"" is the value used to determine when the next low operation event should occur in the synthetic time series. Random draws from this distribution determine when low operation events will occur. To model the magnitude of each event, a transformed and truncated exponential distribution is fitted to the observed low operation capacity factors (Figure 40). The distribution has the following form:

! = !"#$%!%&'() ! 0.8 − !, ! ≤ 0.8 ! = !"# 0, ! > 0.8 where λ is the parameter of the gamma distribution, x is the result of a random draw from that distribution, and y!"# is the resulting magnitude of the low operation event.

The final component to the biomass production model is scheduled outages. Based on personal communication with the operations manager at Fairhaven power, the authors assume that scheduled maintenance occurs twice a year, just before and just after rate period A, which runs from May 1 through October 31. Independent power producers with a qualifying facility contract receive higher prices for their power during period A, so scheduling maintenance around this period maximizes revenue. All biomass power plants in Humboldt County are assumed to use the same strategy in planning schedule outages. However, scheduled outages are not permitted to occur simultaneously, which is a reasonable assumption given that the plants all must coordinate with Pacific Gas & Electric as the local grid coordinator.

The model of scheduled outages is based on observed data, which indicate that on average a single maintenance outage occurs for 7 days. To introduce stochasticity, the duration of each new scheduled outage is modeled as a truncated Poisson distribution with a mean value of 7. The distribution is not truncated in the typical sense, all random draws less than 5 or greater than 14 are discarded instead of being pinned to the extents.

122 With these boundaries on the distribution, the value of the Poisson lambda parameter must be 5.8 for the overall distribution to yield a mean of 7. The model therefore has the following form:

!!"!!" = !"#$$"% 5.8 (redraw if x<5 or x>14)

The final synthetic time series for the biomass production model is generated in the following manner. For each biomass power plant being modeled:

1. Use !!"#$ to generate 365 synthetic data points of the plant at normal operations. 2. Beginning at index 1 do the following:

a. Increment the index by a random value drawn from !!"#$%!!""

b. Replace the value at the current index with a random value drawn from !!"# c. Repeat until the index is greater than or equal to 365.

3. For each biomass plant, draw 2 random values from !!"!!" to determine the length of the schedule outages before and after period A. Set the output of the plants to 0 for the determined number of days as close as possible to the beginning/end of period A but not overlapping any other plant. 4. The times series are all in terms of daily capacity factor. Multiply each series by the plant capacity to get production in units of MW. 5. Repeat each value of the time series 24 times to convert from a daily series to an hourly series 6. Finally, sum the output of all the plants to yield the overall biomass power availability for the year.

123 Figure 38: Observed (Black-Circle) and Synthetic (Red-Triangle) Time Series of Normal Operations at Fairhaven Power Plant

Observed (black-circ) and ARMA(1,1) (red-tri) Series of 'Normal Operation' at Fairhaven 1.00 0.95 0.90 Capacity Factor 0.85 0.80

0 200 400 600 800

Day

Source: SERC Staff (2010)

Figure 39: Distribution of Observed Inter-Arrival Times Between Low Operation Events. Data Modeled Using a Gamma Distribution

Hist - Low Operation Inter-Arrival Days 8 6 4 Frequency 2 0

0 20 40 60 80 100 120 140

Days Between Low Operation Events

Source: SERC Staff (2010)

124 Figure 40: Histogram of the Magnitude of Low Operation Events Data Modeled Using a Transformed Exponential Distribution

Hist - Low Operation (CF<80%) 5 4 3 Frequency 2 1 0

0.3 0.4 0.5 0.6 0.7 0.8

Capacity Factor

Source: SERC Staff (2010)

Table 45 provides additional technical and economic assumptions associated with biomass power production.

D.1.5 Hydropower

The hydropower submodel is a composition of three models that capture separate types of small hydro generation in Humboldt County. There is a small dam in operation at Ruth Reservoir whose operation schedule is unlike any other hydropower installation in the county. The rest of the hydro installations are run-of-river micro-hydro systems whose production patterns are tied to the seasonal rainfall and stream flow. One such micro-hydro system, called Zenia, has a storage pond and therefore operates in a unique manner during the summer months when a premium price is placed on power generated during peak demand. Therefore, the Zenia hydro submodel operates slightly differently than the run-of-river submodel.

D.1.5.1 Ruth Reservoir/Matthew’s Dam Ruth Reservoir has a hydroelectric generation capacity of around 2 MW, with two 1 MW turbines. Monthly generation data are available for July 1983 through July 2009; daily downstream flow is available from the USGS for water years 1980-2009 (USGS station 11480410); the station is immediately below the dam (so these data include both flow captured to generate electricity and flow discharged over the spillway).

125 The model for hydropower production from Matthews Dam is:

P = min(1.226, 0.00524149Q) where P is power generation (MW), and Q is the discharge (cfs). This formula was determined through a two-step process, using the monthly average power generation and discharge data:

1. To calculate the maximum power capacity, average all points with discharge > 650 cfs 2. Choose the slope of the linear section to minimize the sum of the squares of the error.

D.1.5.2 Run-of-River Hydro (Without Peaking)

There is approximately 7.5 MW of installed capacity of run-of-river hydropower in Humboldt County (Table 29). Run-of-river hydroelectric power generation without peaking is modeled on a daily timescale by a seasonal autoregressive model:

#0.92840E't$1 +2.1480Rt Jun $ Nov E't = " !0.98702E't$1 +2.1480Rt Jan $ May, Dec

Et = C min(E't +0.90571, 24)

Where E = daily energy generation (kWh/day) E’ = energy generation, unadjusted, as fraction of capacity R = rainfall (in) C = capacity (kW) t = time (days)

This model was developed using the daily average power generation at Zenia and daily precipitation at the Alder Point Remote Automated Weather Station6. The autoregressive coefficients, coefficient of rainfall, and minimum power production were computed by minimizing the sum of squares of monthly residuals. The model was then scaled according to the power production of each system.

6 Precipitation data available at: http://www.raws.dri.edu/cgi-bin/rawMAIN.pl?caCAPT

126

Table 29: Total Run-of-River Hydropower Generation Capacity in Humboldt County Area Operator City County Capacity (kW) Baker Station Bridgeville Humboldt 1500 Mill and Sulphur Creek Dinsmore Humboldt 995 Tom Benninghoven Garberville Humboldt 25 STS Hydropower (Kekawaka) Alderpoint Trinity 4950

Source: Pacific Gas & Electric (2010b)

D.1.5.3 Run-of-River Hydro (With Peaking) The Zenia system uses water from four creeks, which meet slightly upstream from the powerhouse. Three diversions flow into a storage pond with approximately 12 acre-feet capacity. Gross head is 757 feet from the storage pond to the powerhouse and 887 feet from the Bluford diversion to the powerhouse.

During late summer and early fall, the plant operates only during peak hours (noon to 6:30 PM weekdays); during late spring/early summer, the plant operates at a low level during off-peak hours and peaks at or near full capacity during peak hours.

The model for Zenia is based on the autoregressive model described above with the addition of peaking. Peaking is controlled by an algorithm that models the quantity of water in the storage reservoir. The input for the peaking algorithm is the daily average power generated by the non-peaking model described above. If the input drops below a threshold, a fraction of the input power is stored during off-peak times. At the start of the peak period (noon on weekdays), if the quantity of energy in storage exceeds a threshold, the power output is increased by a fixed amount until the peak period ends or the energy in storage drops below another threshold.

Table 46 provides additional technical and economic assumptions associated with this technology.

D.1.6 Solar Photovoltaic

The solar photovoltaic submodel is based on 15 years of modeled hourly solar radiation data from the National Solar Radiation Database (NREL, 2010). The model is based on data from Station 725945, Arcata Airport, as this is a coastal location in close proximity to Humboldt Bay. This region of the county is home to more than half of the total Humboldt population. Most of the remaining county population lives in areas with less coastal fog; therefore using a coastal data set serves as a conservative estimate of the countywide solar resource. The NREL database classifies the Arcata Airport data set as Class I, which is defined as follows (ibid): If less than 25 percent of the data for the 15-

127 year period of record exceeds an uncertainty of 11 percent, the station receives a Class I designation.

Model application:

• Given a source year (1991-2005) for the solar submodel and a plant capacity (in the case of PV, the plant capacity is the sum of the capacities of every installation in the county): o Determine the amount of solar radiation falling on a surface titled to 41

degrees for every hour of the year using the following formula: !!"#! = !!"#$ ∗

cos ! + !!"## where !!"#! is the solar irradiance on the titled surface, !!"#$ is the irradiance on a surface normal to the sun, ! is the angle of incidence

between the sun and the titled surface in the current hour, and !!"## is the diffuse irradiance (everything except direct sunlight) on a horizontal surface. o For each hour, calculate the output from the entire plant using the following

formula: ! = !!"#! ∗ !!"# ∗ !/!!"# where ! is plant output, !!"# is the plant capacity, ! is the PV derate factor which accounts for all system losses (shading, dirty panels, wire losses, inverter losses, and so forth, a value of 75

percent was used in the model), and !!"# is the solar irradiance corresponding to maximum output of standard PV systems (1000 W/m^2).

Table 47 presents the technical and economic assumptions associated with this technology.

D.2 Balance Support Technologies

D.2.1 Import/Export

The import/export submodel assumes that power can be transmitted in or out of the county any hour of the year. This assumption ignores congestion issues at Humboldt County’s primary points of connection to the greater California grid. These issues are ignored due to the limited scope of this study and because adequate data could not be found to construct a model of congestion.

Table 48 presents the technical and economic assumptions associated with import. D.2.2 Storage

The storage submodel assumes that the storage facility can act as an electric load or generator. The capacities at which the facility charges and discharges are independently specified and can be different magnitudes. Conversion losses are accounted for in the

128 charging, discharging, and storage (in the form of a decay rate) phases of the facility operations according to the following expressions:

Charging: Δ!!!!"#$,! = !!!!"#$,! ∗ !!!!"#$

Discharging: Δ!!"#$!!"#$,! = −!!"#$!!"#$,! ∗ !!"#$!!"!"

Storing: Δ!!"#$%&,! = −!!"#$%&,!!! ∗ !!"#$%&' where Δ!!!!"#$,!, Δ!!"#$!!"#$,!, and Δ!!"#$%&,! are the change in energy stored in hour h due to charging, discharging, and retaining stored energy from the previous hour. !!!!"#$,! and !!"#$!!"#$,! are the rates of charging and discharging respectively in hour h, !!!!"#$ and !!"#$!!"#$ are the conversion efficiencies for charging and discharging, and !!"#$%&' is the hourly decay of energy stored in the facility.

See Table 49 for the technical and economic assumptions associated with this technology. The following sections discuss the three primary storage technologies that were considered by the authors: batteries, pumped hydropower, and compressed air energy storage. D.2.2.1 Batteries Batteries can serve a broad range of applications and are efficient, modular, compact, portable, and capable of providing variable storage capacity. Drawbacks of batteries are high costs, short lifetimes, and the use of toxic materials. Relevant battery chemistries for large-scale renewable energy storage include: lead-acid, nickel-cadmium, lithium ion and sodium-sulfur. While all of these technologies show some promise and have been employed to some extent for utility-scale energy storage, sodium-sulfur (NaS) batteries are likely the best-suited battery technology for supporting large-scale renewables. This technology addresses some of the shortcomings of other battery chemistries such as low cycle lifetime and high cost, and has superior performance characteristics (Figure 41).

129 Figure 41: Comparison of Sodium Sulfur (NAS), Lead-Acid, Lithium-Ion, and Nickel-Metal-Hydride Batteries

Source: NGK (2010)

According to product specifications, NaS modules have a DC efficiency of about 89 percent, incur no self-discharge nor memory effect, and have an estimated 15-year service life depending on depth-of-discharge (2,500 cycles at 100 percent, 4,500 cycles at 90 percent, and 6,500 cycles at 65 percent). Their small footprint, lack of emissions, and insensitivity to ambient temperature makes them easier to site either indoors or outdoors in locations where renewable resources may be in abundance. NaS systems also require little to no regular maintenance. The NaS battery was co-developed by the Tokyo Electric Power Company and NGK Insulators, Ltd. over the course of a decade. It has been demonstrated extensively in its native Japan at over 190 sites, including a 34 MW (245 MWh) facility for wind stabilization (Electricity Storage Association, 2010). As of 2009, NaS batteries represent the most mature energy management capable battery with 270 MW of globally installed capacity (Chi-Jen et al., 2009). NaS storage systems are also gaining traction in the United States. In 2007, American Electric Power, one of the largest investor-owned utilities in the United States, operated a 1.2 MW NaS storage system and was planning to install another twice that size (USA Today, 2007). American Electric Power has since installed other systems in West Virginia, Indiana, Ohio, and was building a 4 MW storage facility in Presidio, Texas in 2010 (National Geographic, 2010). Other major utilities are also pursuing installations of NaS storage systems for peak shaving applications. According to the Electricity Storage Association, annual production capacity planned for 2010 is approximately 150 MW.

130 D.2.2.2 Pumped Hydro The most common type of large energy storage in the world today is pumped hydroelectric energy storage (PHES). PHES facilities generated 4,091 MWh of electricity in the US in 2010 (Energy Information Agency, 2011). The operation of a PHES facility is simple, two reservoirs at different elevations store water as shown in Figure 42. When cheap electricity is available (that is off-peak energy and/or surplus renewable energy) water is pumped from the lower reservoir to the upper reservoir and stored as potential energy. When the stored energy is needed, water is released from the upper reservoir to the lower reservoir through a turbine to generate electricity. The round trip efficiency of a PHES facility is 70 to 80 percent, with a plant life of over 50 years. Costs for PHES are between $1 million to 1.5 million per MW for construction costs and $12-25 per MWh for operations and maintenance costs. Figure 42: Pumped Hydroelectric Energy Storage Schematic

Source: Connelly (2010) While PHES is one of the least expensive forms of energy storage, it is very site dependent. An ideal location would have a large body of water near mountains or hills to provide a large elevation change between reservoirs. A large elevation change is important because the potential energy stored in the water is directly proportional to the difference in elevation between the reservoirs.

There are many hurdles to developing new PHES facilities, including: environmental impacts, costs, long lead times for licensing, lack of transmission lines, and lack of appropriate sites. The most often cited concern for PHES is environmental impacts. Creating a reservoir often requires flooding a large area, which may have negative effects on habitats. Because of this, extensive environmental impact studies are required for new PHES projects and there is often opposition to new facilities. The large capital cost associated with PHES projects is also often a concern for both investors and

131 consumers. For these and other reasons, only one major PHES facility has been built in the United States in the last 15 years (Deane, Gallachoir & McKeogh, 2009) and the Energy Information Agency does not predict much PHES to come online in the near future (EIA, 2011). In Humboldt County one potential site for a PHES facility is adjacent to Ruth Lake on the Mad River. A big advantage of siting a facility adjacent to Ruth Lake is that it could take advantage of existing infrastructure, including using the lake itself as the lower reservoir and using the existing transmission lines and interconnection facilities that serve the existing 2 MW hydroelectric facility located at the lake. As shown in Figure 43, there are two locations near Ruth Lake that exhibit potential for an upper reservoir. In either location an impoundment would need to be constructed and a penstock and powerhouse would need to be installed. Initial estimates for the capacity indicate that each reservoir has the potential to store enough water to power a 20 MW generator for more than two weeks continuously. Figure 43: Potential Upper Reservoir Sites Near Ruth Lake.

Source: Generated by Schatz Energy Research Center staff using Google maps.

D.2.2.3 Estimation of Pumped Hydro Upper Reservoir Capacity

To determine the volume of each potential reservoir as identified in Figure 43, Google Earth was employed with a topographic map overlay. For each location, a logical site was chosen for an impoundment that would create the reservoir. The maximum water elevation of reservoir 1 located in area 1 was chosen to be 3600 feet, and reservoir 2 in area 2 had a maximum elevation of 3360 feet (Figure 44).

132 Figure 44: Reservoirs Outlined on Topographic Map With Impoundment Locations

Source: Google Earth (2010)

For each reservoir, the authors estimated the surface area of a plane intersecting the reservoir at the contours in the topographic map using the area tool in Google Earth (Figure 45 and Figure 46). Once all of the surface areas of the contours were estimated, numerical integration was used to calculate the total volume based on the following equation:

!!!

! = !"#$! ×!"#$ℎ !!! where V is the volume of the reservoir (ft3)

n is the number of contour areas for the reservoir

Area is the area estimated for the contour slice (ft2)

Depth is the distance between contour lines, for this map 80 feet.

133

Figure 45: One Area Slice in Reservoir 1

Source: Google Earth (2010)

Figure 46: Head-On View of All Area Slices in Reservoir.

Source: Google Earth (2010)

Reservoir 1 has an estimated volume of 134,000 Acre-feet (165 million cubic meters); reservoir 2 has an estimated volume of 132,000 Acre-feet (162 million cubic meters). The

134 authors performed several calculations for each reservoir to determine the number of days of storage for each reservoir. The scenarios investigated for each reservoir are:

• Continuous generation for a 2, 5, 10 and 20 MW generator. • Head losses at the generator equal to 0, 20 and 50 percent of the static head.

For each reservoir, the authors assumed a maximum drawdown of 80 feet, a figure within the range of current PHES facilities – the Bath County Pumped Hydro Storage Station in Virginia has an upper reservoir that fluctuates up to 105 feet – and convenient because it is the difference between the contours lines of the topographic map. The amount of time of generation is estimated with the following equation: !×!×!×ℎ×! ! = !

where t is the time of full power generation (sec)

V is the volume of water used to generate power (m3)

ρ is the density of water (kg/m3)

g is the gravitational constant (9.8 m/s2)

h is the change in elevation of the water (m)

η is the efficiency of the system, includes head loss and efficiency of turbine

P is power (w).

Each reservoir has enough capacity, even with a 50 percent loss in static head, to generate for 27 days at 20 MW, see results in Table 30. Because the capacity is large, the size of the reservoir could potentially be reduced to decrease construction costs and environmental impacts. If the reservoir height was limited to 3360 feet for reservoir 1 and 3200 feet for reservoir 2, there would still be enough capacity for 2 weeks continuous generation at 20 MW with a 50 percent loss in static head. Table 30: Days of Continuous Generation With the Given Generator Size and Head Loss

Generator 2 MW 5MW 10MW 20MW

Head loss 0% 20% 50% 0% 20% 50% 0% 20% 50% 0% 20% 50%

Reservoir 1 595 476 297 238 190 119 119 95 59 59 57 29

Reservoir 2 558 446 279 223 178 111 111 85 55 55 44 27

Source: SERC Staff (2010)

135 D.2.2.4 Compressed Air Energy Storage Compressed Air Energy Storage (CAES) is a technology that can be used for storing large amounts off-peak and/or surplus renewable energy. Surplus electricity is used to compress air into an underground storage medium. When electricity is needed, the compressed air is heated and expanded in a turbine to generate electricity. The technology is currently commercially available at utility-scales and can provide long- term energy storage (on the order of hours to a few days). The technology has fast ramp up rates and good partial load efficiency. The operation of a CAES facility is very similar to a conventional gas turbine (Brayton cycle), except the compression and expansions cycles are independent and happen at different times. Air is compressed using cheap energy (for example off peak and/or surplus renewable generation) and stored in a reservoir under pressure (20-60 bar, 290- 870 psi). When the stored energy is needed the air is heated (typically using natural gas but other fuels can be used) and expanded in a generator as shown in Figure 47 (Succar & Williams, 2008). Defining a single performance index for CAES is difficult because of the different inputs used: electricity to drive the compressors and natural gas or other fuel to heat the air during expansion. While there are many ways proposed for quantifying the performance (for example, a thorough discussion is given in (Succar & Willams, 2008, pgs. 36-41)) the most basic way to define efficiency is based on the ratio of electricity generated by the turbine (ET) to the sum of the electricity used by the compressors (EC) and the natural gas used during expansion (ENG).

!! ! = !! + !!" Using this formulation with typical values from current CAES facilities, the efficiency is approximately 54 percent. A major drawback to CAES is the need for additional fuel in the expansion process. There are schemes for minimizing, and potentially eliminating this need by utilizing heat storage for the heat created during the compression process. This type of advanced adiabatic CAES (AA-CAES) system has not been built and as of yet is unproven (Electric Power Research Institute, 2003).

136 Figure 47: Typical CAES System Configuration

Source: Succar & Willams, 2008

Particular geologies are required for underground air storage. The geologies that are potentially suitable for a CAES facility are salt caverns, hard rock, and porous rock or aquifers. Smaller systems (approximately 10 MW) can also be built using pipes, either above ground or below, as the storage medium. Extensive geological studies must be done to determine whether a site meets the necessary requirements. Costs for building a CAES project vary depending on the geology. The Electric Power Research Institute (EPRI) gives cost estimates for a CAES project with an assumed plant life of 20 years (Table 31).

137 Table 31: Cost Per kWh for Various CAES Geologies

Geology Cost ($/kWh 2003 dollars) Solution mined Salt cavern 1 Dry Mined Salt Cavern 10 Underground rock caverns 30 Porous rock formations 0.10 Abandoned mine 10

Source: EPRI, 2003

There are currently two commercial CAES plants in the world, one in Germany and one in the United States. The Huntorf plant in Bremen, Germany, came online in 1978 and was designed to provide black start services to nuclear power plants and serve as a peaking power plant. It is now also being used to help balance an increasing wind load from Northern Germany. It has a maximum power output of 290 MW and can provide full power for three hours. The storage consists of two salt caverns with a total volume of 310,000 m3 with operating pressures ranging from 48 bar to 66 bar (approximately 700 psig to 950 psig). The plant has operated reliably for three decades with availability and starting reliability of 90 percent and 99 percent, respectively. The McIntosh plant in Alabama began operation in 1981. It has a power output of 110 MW and can provide full power for 26 hours. It uses a single 560,000 m3 salt cavern as storage, with operating pressures ranging from 45 bar to 75 bar (approximately 650psig to 1,100 psig). After some initial startup problems the plant has had over 10 years of reliable operation with starting and running reliabilities over 90 percent. The McIntosh plant utilizes a heat recuperator that reduces fuel use by approximately 22 percent compared to the Huntorf plant. The operation parameters for McIntosh are similar to the Huntorf plant. In addition to the two commercial plants, there are several CAES plants being planned. The largest is the Iowa Stored Energy Project. The proposed $400 million, 270 MW plant would utilize power from a 75-150 MW wind farm, in addition to cheap off peak electricity, to charge a sandstone aquifer. Service is expected to begin in 2015 (Iowa Stored Energy Project, 2011). Pacific Gas and Electric is planning a 300 MW CAES plant in Kern County. In September 2009 they were awarded $24.9 million in matching funds from the California Public Utilities Commission for the project, which is currently in the early stages of planning. Underground saline rock formations would provide storage for up to ten hours of full generation (California Public Utilities Commission, 2010). In Humboldt County there exists a potential for CAES in closed natural gas wells. There are many natural gas wells in the Eel River Basin that are currently not under

138 production that could be suitable sites for CAES. Extensive geologic testing would need to be performed to determine if the sites are suitable for development. Important parameters to assess are the permeability and porosity for the desired location. Estimated costs for a CAES system in the Eel River Basin may fall in the $0.10 per kWh range (Table 31) as the geology in this region is characterized as porous rock. Residual hydrocarbons may pose a potential problem for development of closed gas wells. Residual hydrocarbons could produce permeability-reducing compounds and corrosive materials when exposed to air. Furthermore, there could be the potential for combustion when high-pressure air is introduced in the presence of residual hydrocarbons. However, there are mitigation techniques that can be employed to reduce these risks (Succar & Willams, 2008). D.2.3 Reserve Capacity

To model the reserve capacity required during any hour of the year, the authors conducted an analysis of the uncertainty associated with hour-ahead forecasts of demand, wind power production, and wave power production. The forecast method is a persistence forecast, meaning the demand or intermittent power production in the next hour is assumed to be equal to that in the current hour. In reality, grid operators can employ a multitude of forecasting techniques that produce results far more accurate than persistence forecast. However, a persistence forecast is employed as a highly conservative measure of the uncertainty associated with demand and production.

Using two coincident years of hourly data for demand, wind, and wave production in Humboldt County, first difference between every hour and the preceding hour is calculated yielding three vectors of forecast residuals. These vectors are analyzed to determine the overall variation in forecast error and how it depends upon the season, the time of day, and the magnitude of overall demand, wind, and wave capacity.

Ultimately, the reserve requirement is calculated via the following steps: 1) calculate the combined variance of the forecast error for all three vectors, 2) take the square root to get the standard deviation, 3) multiply by three. Assuming a normal distribution of forecast residuals, this three sigma reserve requirement means that reserve capacity will be adequate approximately 99.9 percent of the time. Using data specific to Humboldt County’s electricity demand and wind and wave resource, the authors have found that the average reserve requirement with no renewables (demand alone) is equal to 12.7 percent of peak demand. The average reserve requirement if wind and wave alone meet 75 percent of load on an energy basis is 30.5 percent of the peak demand. These reserve requirements would be much lower if a more sophisticated forecasting method were chosen.

After running a sensitivity analysis that included various combinations of wind and wave generation that reached 250 MW combined and demand growth that reached a 165 MW annual average, it was determined that the 160 MW natural gas plant was always able to satisfy the reserve requirement without curtailing any of the other generation

139 sources. However, there were frequent occurrences when the reserve requirement meant that electricity import was necessary to satisfy the local demand.

D.3 Demand Submodels

D.3.1 Population Growth

Population growth estimates were used to model the expected growth in demand for energy services. The population submodel is a simple exponential growth model that projects future population based on the U.S. Bureau of the Census 2010 estimate for Humboldt County and an assumed growth rate according to the following formula:

where !! is the future population in ! years, !! is the present population, and ! is the growth rate. The default growth rate (0.36 percent) is based on a study by Auffhammer and Aroonruengsawat (2010), which analyzed the impact of population, price, and climate on future residential energy consumption. They investigate three population growth scenarios for the state of California: 0.18 percent, 0.88 percent, and 1.45 percent. Their growth rates are based on data from the California Department of Finance (CA DOF, 2007). According to Department of Finance data, Humboldt County population is expected to grow at a rate 41 percent lower than the state average. So the three growth scenarios from the Auffhamer study are scaled to produce Humboldt-specific low, medium, and high growth rate cases: 0.07 percent, 0.36 percent, and 0.59 percent.

Given a projected population for Humboldt County, the question still remains whether there may also be a change over time in the per capita demand for energy services. For example, new appliances and electronic devices can create new demand for services that did not previously exist. Conversely, cultural changes resulting in an increased conservation ethic can reduce demand. Due to the constrained scope of this study, the authors simply assume that future per capita demand will be equivalent to present values. Therefore, to model growth in energy demand, demand is scaled linearly with population growth.

D.3.2 Humboldt Natural Gas and Electricity Demand D.3.2.1 Baseline Electricity and Natural Gas Use The estimate for baseline hourly electricity demand is based on a data set received from PG&E describing the whole-county electric load at 15-minute intervals from Jan 1, 2004 to December 31, 2008; it contains five years of data (PG&E 2010c). These data do not include on-site loads at industrial facilities that also produce on-site power (for example, some saw mills and pulp mill). The authors created an hourly countywide dataset by replacing any missing values and finding the average electricity use in each hour. The data from 2007 and 2008 are the best quality in the dataset in terms of having few missing values and having sector-level information in those years, so these years are used as the basis for further estimates of hourly electricity demand for various scenarios.

140 Natural gas is the most commonly used heating fuel in Humboldt County and is the best documented. The authors obtained a monthly natural gas data set parallel to the electricity data from 2004-2008.

Table 32 below lists the sectors and subsectors for which the authors arrived at estimates of electricity and natural gas consumption. In some cases there is not an exact match between the sector definitions for the two energy types. Overall, the residential sector dominates both electricity and natural gas demand in the county.

Depending on the scenario of interest, the demand algorithm adds and subtracts load shapes from the annual hourly data to modify its magnitude. A load shape is 8760 hourly values with energy units that can be added to the baseline load to modify it. If the values in the load shape are positive, the modified load in that hour will be higher than the baseline and vice versa.

141 Table 32: Electricity and Natural Gas Consumption in 2008 for Humboldt County

Percent of Percent of Electricity Number of Natural Gas Number of Countywide Countywide Electricity Sector Name Consumption Customers Natural Gas Sector Name Consumption Customers Consumption Consumption (MWh/year) (#/year) (therms/year) (#/year) (%) (%) Residential 448,202 49% 56,353 Residential (Total) 22,463,924 56% 40,715 Commercial Buildings Commercial Buildings 238,151 26% 4,248 8,600,807 22% 2,305 (disaggregated below) (disaggregated below) College 12,637 1% 51 College 1,357,227 3% 40 Food / Liquor 37,619 4% 145 Food / Liquor 366,175 1% 72 Health Care 18,219 2% 117 Health Care 1,129,455 3% 96 Hotel 11,733 1% 143 Hotel 908,970 2% 76 Warehouse 9,637 1% 271 Warehouse 108,465 0% 112 Office 40,621 4% 1,209 Office 1,023,157 3% 685 Refrigerated Warehouse 2,895 0% 22 Refrigerated Warehouse 5,901 0% 10 Restaurant 21,749 2% 340 Restaurant 1,118,732 3% 213 Retail Store 27,409 3% 559 Retail Store 374,410 1% 338 Schools 10,193 1% 172 Schools 576,920 1% 97 Miscellaneous 45,439 5% 1,218 Miscellaneous (includes 1,631,395 0 567 Commercial (Other) 26,978 3% 723 any other Commercial) Street lighting 4,367 0% 1,137

TCU 98,058 0% 66

Industrial 117,496 13% 290 Industrial 6,366,366 1% 113 Agriculture 31,297 3% 914 Agriculture and Water 1,785,933 9 Water Pumping 11,460 1% 158 Pumping Mining and Mining (no construction 1,185 0% 78 38,638 0% 29 Construction included) Unclassified 26,421 3% 2,501 Unclassified 630,398 2% 717 TOTAL 905,557 100% 66,402 TOTAL 39,984,124 100% 43,956

Source: CEC (2010a)

142 D.3.2.2 Heat Demand Estimates The authors estimated the hourly demand for thermal energy7 in Humboldt County buildings using a combination of top-down and bottom-up approaches. The top-down approach results in an overall estimate for heat demand on a monthly basis from natural gas sales data and census data, corrected for weather. The bottom-up approach estimates the relative hourly distribution of heat demand and the total space cooling demand in commercial buildings (cooling in residential buildings is essentially nonexistent) with DOE2 building energy models from the Energy Commission Database for Energy Efficient Resources (DEER). The combination of the top-down and bottom-up estimates results in an hourly profile for heating and cooling demand

The demand for thermal energy is ultimately driven by peoples’ desire to change the temperature of their indoor environment, food, and water. Space conditioning is a dominant driver for thermal demand for heating and cooling, but domestic hot water, food preparation, and refrigeration are also significant. In specific cases, like district heating, losses in heating distribution systems can be a significant part of the load; there are not many systems of that type in Humboldt County.

Space heating and cooling demands are complex; they depend on the physical and operational features of buildings, weather, and interaction with the building occupants. However, given a relatively fixed fleet of buildings and occupants, the day to day changes in thermal demand can essentially be attributed to four meteorological factors: outside air temperature, humidity, solar radiation, and wind speed (Wojdyga, 2008; Heller, 2002; Enshen, 2005). Of those, air temperature is typically the strongest driver for thermal demands and is often referred to in terms of degree-days– defined as the difference between the daily average temperature and some neutral reference point, traditionally 65°F. Days with average temperatures above the reference point are counted as cooling degree-days (CDD) and those with temperatures below are counted as heating degree-days (HDD), for example, a day with an average temperature of 55°F is counted as 10 HDD. Typical weather patterns in Humboldt County8 result in approximately 4,500 HDD and 0 CDD each year; Figure 48 shows the 30-year average values for Humboldt County in the context of the population-weighted distribution for cooling and heating degree-days in the United States as a whole. Humboldt County is fairly typical in terms of HDD but is on the extreme low end for CDD. There is very little need for active space cooling in Humboldt County because the population mostly lives in the mild coastal zone.

7 Unlike many places, there is very little demand for space cooling in Humboldt County – another thermal demand component – although refrigeration for cold storage is similar to other places. Those refrigeration demands are considered weather independent for our study and are included in basic electricity demand estimates.

8 The weather data presented here are from the Arcata / Eureka Municipal Airport, which is sited in a coastal area. While parts of the County experience much more varied (hotter and colder) weather than the airport, the majority of the County’s population and buildings are located in the coastal zone.

143 Figure 48: Annual Cooling and Heating Degree Days; Nationwide Distributions (RECS 2005) and Humboldt County 30-Year Averages (NOAA 2010)

Source: SERC Staff (2010)

D.3.2.3 Residential Buildings in Humboldt County The residential buildings sector is the best-characterized group in Humboldt County thanks to the decennial U.S. Census, which includes survey questions about housing stock. The approximately 130,000 residents of the county live in 60,000 housing units in 2010 – an average density of 2.2 persons per household, which is below the California average of 2.5 persons per household. Based on Census 2000 estimates, nearly 70 percent of the housing stock is single- family detached housing; 10 percent is mobile homes. The full estimate of housing stock distribution is included in Table 33.

144 Table 33: Humboldt County Housing Stock Type of Structure Number Percent 1-unit, detached 38,293 69% 1-unit, attached 1,542 3% 2 units 2,218 4% 3 or 4 units 3,387 6% 5 to 9 units 2,096 4% 10 to 19 units 1,160 2% 20 or more units 1,282 2% Mobile home 5,481 10% Boat, RV, van, etc. 453 1%

Source: US Census 2000

A housing boom occurred in parallel with a growing timber industry in the middle of the 20th century. The increased building in that period shows prominently in Figure 49, which shows the distribution in building age from the 1990 and 2000 Census. In 2000, 74 percent of the housing in use was constructed before 1980, when California Title 24 building codes were put into effect. The building codes have been very successful since then at reducing space conditioning demands, but relatively few Humboldt County homes were built under Title 24 codes. As new homes are built and old ones vacated, the energy efficiency of the residential housing fleet will continue to improve.

Figure 49: Age Distribution of Humboldt County Building Stock

Number of Housing Units! Year of Construction! 0! 5000! 10000! 15000! 20000!

1999 to March 2000! 1995 to 1998! 1990 Census! 1990 to 1994! 2000 Census! 1980 to 1989! 1970 to 1979! 1960 to 1969! 1940 to 1959! 1939 or earlier!

Source: US Census 2000

D.3.2.4 Residential Heating Fuels

Table 34 shows the diversity of heating fuels used in California households; the data are from the RECS 2005 Microdata dataset produced by the Energy Information Administration. It also

145 includes results from the 2000 census for the primary space heating fuel of Humboldt County households. Of 468 California households that were surveyed, 408 (87 percent) had access to utility natural gas and the other 60 (13 percent) did not; the households are grouped according to utility gas access. The fraction of Humboldt County Households with access to utility- delivered natural gas in 2007 was 76 percent (from Energy Commission 2010a and Census 2000 data). The data show that those with access to natural gas overwhelmingly use it for water heating (96 percent) and large majorities also use it for space heating (80 percent) and cooking (66 percent). The percentage of Humboldt County residents who reported using natural gas as their primary heating fuel in 2000 was 62 percent, which is in line with the 61 percent estimate one would arrive at by multiplying the percentage of households with access (76 percent) by the typical fraction who use gas for heating, given they have access, in California (80 percent).

Those who lack access to utility gas must turn to other fuels for their thermal needs. In this case electricity is the clear market leader in spite of the typically high cost of electric water heating and cooking (and space heating if resistive coils are used). Californians report that they use wood for space heating at similar levels regardless of their access to utility gas; the median annual use regardless of utility gas access is 9 cords. The proportion of Humboldt County residents who use wood for their primary heating fuel is higher, 18 percent, but has fallen since 1990 from 27 percent and may be continuing to fall.

Overall, the data show that it is a good assumption that people in California who use utility gas for space heating also use it for water heating and likely use it for cooking.

Based on these data, the authors make the assumptions summarized in Table 35 for Humboldt County household water heating and cooking fuel utilization based on their space heating fuel. The space heating fuel market share estimates (shown in Figure 50) are an average of the 2000 Census results and a linear estimate for the 2010 results (that is, they are equal to the sum of the 2000 Census result and ½ the difference between the 1990 and 2000 results). The assumptions for domestic hot water and cooking are based on the conditional probability of a household in California using a particular water heating or cooking fuel given their primary space heating fuel, and are rounded to the nearest 10 percent.

146 Table 34: Primary Space Heating, Water Heating, and Cooking Fuels Used by California Households Has Access to No Access to Humboldt Utility Utility County Natural Gas Natural Gas Census (n=408) (n=60) 2000

Primary Space Electricity 16% 70% 10% Heating Fuel Natural Gas 80% 0% 62%

LPG 1% 15% 8%

Fuel Oil 0% 5% 2%

Wood 2% 5% 18%

None 1% 7% 0.3%

Primary Water Electricity 3% 67% Heating Fuel Natural Gas 96% 0%

LPG 0.2% 32%

Solar 0.2% 2%

Primary Electricity 31% 80% Cooking Fuel Natural Gas 66% 0%

LPG 0% 15%

None 3% 5%

Data grouped by access to utility gas service shown; also shown for reference are Humboldt County households in general.

Source: RECS 2005 Microdata and US Census 2000

147 Table 35: Residential Sector Heating Fuel Market Shares and Conditional Water Heating and Cooking Fuel Market Shares Based on Space Heating Fuel Space Heating Fuel Water Heating Fuel Cooking Fuel

Natural Gas Natural Gas – 100% Natural Gas – 70%

(64%) Electricity – 30%

Wood Natural Gas – 70% Natural Gas – 40%

(13%) Electricity – 10% Electricity – 40%

LPG – 20% LPG – 20%

Electricity Natural Gas – 60% Natural Gas – 30%

(11%) Electricity – 40% Electricity – 70%

LPG Natural Gas – 15% Natural Gas – 10%

(10%) LPG – 85% Electricity – 30%

LPG – 60%

Fuel Oil Electricity – 100% Electricity – 100%

(1%)

Other Natural Gas – 50% Natural Gas – 50%

(1 %) Electricity – 20% Electricity – 30%

LPG – 30% LPG – 20%

Note: Baseline Humboldt County 2008 estimates.

Source: SERC Staff

148 Figure 50: Humboldt County Heating Fuel Market Shares in 1990, 2000, and Estimated for 2010.

70! 60! ! 1990! ! 50! 2000! 40! 2010 (est.)! 30! 20! Households

using fuel (%) 10! 0! ! ! ! ! ! ! ! ! !

Wood Utility gas Electricity Other fuel No fuelSolar used energyCoal or coke

Bottled, Fueltank, oil, or kerosene,LP gas etc.

Source: US Census 1990,2000 and SERC Staff

D.3.2.5 Residential Heating Technologies

Domestic Hot Water

Domestic hot water accounts for 67 percent of the non-weather related natural gas demand for California households who use natural gas to cook, heat water, and heat their home. The authors assume the same allocation for Humboldt County households that meet those criteria.

Typical Domestic Hot Water (DHW) performance (for example, the performance of a gas- powered water heater with a tank) is described in terms of energy factor (EF), which combines the efficiency of a burner with standby losses (Energy Commission, 2008). The 2008 Update to DEER describes an analysis of California tank water heaters. For the purposes of this study, the assumptions for the baseline water heater technology are summarized in Table 36. It is assumed that the baseline EF of water heaters in Humboldt County is evenly distributed among those that were identified as the typical water heater that was produced between 1993 and 2008. The burner efficiency is assumed to be the average burner efficiency among those identified in the California small gas water heater database (California Energy Commission 2008). The standby efficiency is inferred from the other two estimates. For electric water heaters, the energy factor is equal to the standby efficiency – 71 percent for the baseline estimate.

New construction and replacements of DHW will have to be in compliance with Energy Commission appliance standards, which specify a minimum EF of 0.62.

149 Table 36: Baseline Humboldt County Gas Water Heater Efficiency Energy Factor Burner Efficiency Standby Efficiency

56% 78% 71%

Source: SERC Staff based on Energy Commission 2008

Home Heating Equipment

Gas furnace efficiency assumptions are based on the DEER documentation for 2005 (Itron 2005) and 2008 (California Energy Commission 2008). They are summarized in Table 37. Table 37: Baseline Humboldt County Gas Heating Equipment Efficiency Single Multifamily Mobile Weighted Family Home Total

AFUE (%) 78% 73% 78% 77%

Humboldt Co. Share (%) 69% 20% 11% 100%

Source: Itron 2005 and Energy Commission 2008

Cooking

Cooking accounts for 33 percent of non-weather related natural gas demands (according to analysis of RECS 2005 data for California households). The estimates of cooking efficiency are based on Lawrence Berkeley National Laboratory’s Home Energy Saver documentation (Mills et al, 2008). For this analysis, the heat rate of electric ranges and ovens is assumed to be the theoretical lowest amount of heat that can be used to cook (culinary concerns and induction technology aside, among other simplifying assumptions). Therefore, the efficiency of natural gas cooktops or ovens is equal to the electricity requirements divided by the natural gas required for a similar appliance. The authors take the baseline assumptions about natural gas and electric stove/oven combinations from Mills et al, 2008 to make the baseline assumption presented in Table 38 below. The main conclusion is that the heat rate for natural gas cooking is 2.2 times higher than electric cooking; that same factor is assumed for cooking with LPG.

150 Table 38: Cooking Appliance Baseline Assumptions Cooking Appliance Fuel Electric Natural Gas

Oven heat rate (kBtu/hr) [efficiency] 7.847 [100%] 11 [71%]

Oven hours of use per week 2 2

Stove heat rate (kBtu/hr) [efficiency] 3.412 [100%] 9 [38%]

Stove hours of use per week 7 7

Time-weighted efficiency 100% 45%

Source: Mills et al 2008

D.3.2.6 Commercial Buildings

The authors estimate that for most commercial sub-sectors, the proportion of buildings with access to natural gas is roughly equivalent to the proportion of Humboldt County’s population with access (75 percent) and that every commercial building with access will use natural gas for heating, cooking, and hot water preparation. Based on local knowledge, 100 percent of the healthcare and college facilities are assumed within reach of natural gas (while there are exceptions, this rule is generally true in the county). Of the buildings without access to natural gas heat, trends are used from CBECS 2003 microdata for the Pacific Census Division (California, Oregon, and Washington). Table 39 below presents those trends for fuel utilization: Table 39: Share of Heating Fuel By End-Use for Commercial Buildings in CBECS (2003), Pacific Census Division, Without Access to Utility Natural Gas (n=126). Fuel LPG Electricity Fuel Oil Heating Share (%) 4% 79% 17% Hot Water Share (%) 5% 78% 18% Cooking Share (%) 4% 80% 17%

Source: CBECS 2003

D.3.2.7 Commercial Building Heating Technologies Commercial space heating equipment in Humboldt County is often different from the equipment used in typical commercial buildings around the state of California. Anecdotally, commercial buildings in Humboldt County tend to be smaller and rely on technology that is typically found in residential construction. For instance, the authors are aware of at least three large commercial facilities in the area use multiple residential furnaces and air conditioning units to serve the building rather than a centralized system. This design practice leads to lower

151 efficiency than is typically possible. The trend is likely due to a lack of capacity in the area for best-practice commercial HVAC design and installation.

To account for the residential nature of many commercial buildings in Humboldt County, some of the baseline assumptions from the residential sector are applied to the commercial sector where appropriate.

Space Heating

Commercial space heating equipment in Humboldt County will be assumed to be equal to the equipment used to generate the baseline DEER 2005 – meaning an assumption of 80 percent thermal efficiency for gas heating equipment.

D.3.3 Heat Demand Algorithm

The following two steps are taken to develop hourly electricity demand for heating and monthly demand for various fuels:

1. Estimate Heat Demand: find the hourly demand for heating service across the county in each sector. 2. Apply Technology: use a fleet of heating technologies to meet the heating end-use service demand in each hour; develop a time series demand for fuels and electricity.

D.3.3.1 Estimate Heat Demand The approach used to estimate countywide heat demand in commercial and residential buildings was to use monthly natural gas consumption to estimate the demand for heat in each sector on an end-use basis, accounting for the typical efficiency of equipment, market share (that is, percent of buildings in the sector who use natural gas as a heating fuel), and using an estimate of the fraction of the total natural gas that is used in each end-use category. In general, a month of natural gas use was disaggregated into end-use heat demands by the following approaches.

The estimates for heat demand in each sector began with monthly-billed natural gas data for a five-year period. The authors used a linear model for degree-days to estimate the split of natural gas consumption between space heating (weather dependent) and relatively weather- independent uses, namely domestic hot water (DHW) preparation and cooking. The basic equation for estimate space heating demand is presented below.

Equation 1: % 12 ( "HVAC ' $QNG # Qconst * & month ) QHVAC load = + HVAC,NG

!

152 where:

QHVAC load, yearly = the annual space heating load in a given sector

QNG = the monthly natural gas use

Qconst = the monthly weather independent natural gas use

φHVAC,NG = the percentage of buildings in the sector that use natural gas heat

ηHVAC = the average thermal efficiency of natural gas heating systems

The term Qconst referenced in the equation above can be estimated by plotting HDD versus heating fuel consumption by month for a given sector, then using a linear estimate to find the fuel consumption when there are no heating degree days in a month. A critical factor in the analysis is the choice of baseline temperature for HDD estimates. The approach used by the authors is a combination of trial and error and practical knowledge of Humboldt county buildings.

Figure 51 shows how the choice of HDD base temperature affects the inference that can be made about relative fractions of space heat and DWH/cooking. Each row of plots uses a different base temperature (55°, 57.5°, and 60° F); the left column shows therms of gas in the residential sector versus HDD/month and the right shows a time-series of the gas consumption data on the same ordinate scale.

Figure 52 shows that among the RECS households, there was a relationship between the estimated fraction of natural gas use for heating and the annual heating degree-days. The plot includes a simple linear fit for the data, which is not a very good model for the data but provides some guidance about the general trend. If the homes in Humboldt County follow the general trend, space heating will represent about 60 percent of the natural gas use in residences (at approximately 5000 HDD/year). Based on that trend and anecdotal knowledge that there is some small amount of heating even in the summer in Humboldt County, the authors selected 57.5 F as a base temperature for heating in the residential sector. There was not as exhaustive a ground-truthing for the estimates in other sectors, which account for much smaller fractions of the overall gas use.

153 Figure 51: Residential Natural Gas Consumption Disaggregated Into Weather-Related (Space Heat) and Non-Related (Constant) Parts Using Three Different Base-Temperatures for Heating Degree Days

5HVLGHQWLDO1*8VH+''EDVHï Spaceheat = 48% Constant (DHW, cooking, etc.) = 52% 4e+06 4e+06 2e+06 2e+06 0e+00 0e+00 Residential Natural Gas Use (therms) 0 200 400 600 800 Residential Natural Gas Use (therms) 2004 2005 2006 2007 2008 2009

HDD/month time UHVLGHQWLDO1*8VH+''EDVHï Spaceheat = 63% Constant (DHW, cooking, etc.) = 37% 4e+06 4e+06 2e+06 2e+06 0e+00 0e+00 residential Natural Gas Use (therms) 0 200 400 600 800 residential Natural Gas Use (therms) 2004 2005 2006 2007 2008 2009

HDD/month time

5HVLGHQWLDO1*8VH+''EDVHï5HVLGHQWLDO1*8VH+''EDVHï 6SDFHKHDWSpaceheat &RQVWDQW '+:FRRNLQJHWF = 79% Constant (DHW, cooking, etc.) =ï 21% 4e+06 4e+06 4e+06 4e+06 2e+06 2e+06 2e+06 2e+06 0e+00 0e+00 0e+00 0e+00 Residential Natural Gas Use (therms) Residential Natural Gas Use (therms) Residential Natural Gas Use (therms) 0 200 400 600 800 Residential Natural Gas Use (therms) 20042004 20052005 20062006 20072007 20082008 20092009

HDD/month timetime

Source: SERC Staff

154 Figure 52: Fraction of Natural Gas Attributed to Space Heating Versus Yearly Heating Degree Days (Base Temperature 65°F). 1.0 0.8 0.6 0.4 0.2 0.0

Fraction of Natural Gas devoted to Space Heating 0 2000 4000 6000 8000 10000

Heating Degree Days (HDD65)

Source: SERC Staff based on RECS 2005 Microdata

With monthly natural gas data split into weather-related and non-related parts for each sector, there was a further breakdown of the non-related parts into cooking and hot water preparation. Those fractions were informed by assumptions put forth in Itron 2005, CBECS, and others. With an assumption that the natural gas users are representative of all the heating consumers in the county in terms of demand levels, Equation 2 is used to estimate the total demand for heating energy service on a monthly basis in each sector, and end-use category. The result of the calculation is an estimate of the demand for heat (in GJ) in each.

Equation 2

C f , j,k ! Ff ,i, j !! f ,i EUIi, j,k = S f ,i ! S f ,k

where:

f = fuel “f” i = end-use “i” j = month “j” k = sector “k” EUI = end-use heat intensity (Joules)

155 C = consumption of fuel f in month j F = fraction of consumption of fuel f attributed to the end-use i given 100 percent market share; may be a function of month j. ηf,j = efficiency of end-use for fuel j in sector j Sf,i = market share of fuel f for the end-use technology i Sf,k = market share of fuel f among the sector k.

The monthly heating demands (now in terms of sector and end-use) are broken down further by applying hourly load shapes. The load shapes are drawn from Itron 2005 and in some cases (that is, for residential and commercial buildings) are customized for Humboldt County weather in 2008 using the basic DOE2 energy models that informed DEER originally.

The resultant hourly loads are for the baseline fleet of buildings and heat end-users in the county and are assumed to be applicable to future buildings as well (that is, there is no provision built into the model for heating load reduction from weatherization programs, and so forth)

D.3.3.2 Apply Technology A fleet of technology is deployed to meet the hourly demand for heating service in each sector. The technologies can be baseline-only or include a mix of baseline and updated technology (for example, switching from natural gas furnaces to electric heat pumps or using a more efficient furnace). To arrive at estimates for fuel or electricity demand in each hour, Equation 2 is recast in terms of Cf,j,k as shown below: EUI ! S C = i, j,k f ,k f , j,k ! f ,i

Using this equation, the energy use intensity in each hour is scaled by market share for each fuel and the efficiency of the conversion technology. The demands for each fuel are summed across the sectors to arrive at an hourly demand for the diverse fuels used to provide heating services to Humboldt County homes, businesses and other facilities.

D.4 Energy Balance Algorithm

The energy balance algorithm reconciles the demand with the available generation according to specified priorities in every hour of the year while satisfying reserve requirements. An energy storage strategy is implemented as part of the energy balance algorithm.

A simplified formulation of the algorithm follows:

1. Always meet reserve requirements. 2. Meet demand with available supply of priority. 3. Export excess power from generators whose priority is below the export priority threshold.

156 The reserve requirements are based on the reserve submodel described in Appendix D.2.3. Each generator has a reserve fraction parameter that indicates the fraction of the generator’s unused capacity that be counted toward the reserve requirement for a given hour.

The following pseudo code details the steps followed in the energy balance algorithm to reconcile all of the demands and priorities required for power dispatch. Table 40 provides example parameter values for generators; these are helpful to decipher how the mechanics of the algorithm dispatch power within the constraints on the grid. Table 40: Example Parameter Values Helpful for Illustrating the Energy Balance Algorithm Generator Capacity Priority Stability Reserve Fraction Min Share Turndown

Renewables 40 1 0 (1 for 0 (1 for hydro) 0 (wind, wave, etc.) hydro)

Biomass 60 2 1 1 10

Storage 15 3 0.2 0.2 5

Gas 150 50 1 1 0

Import 60 99 1 1 0

Export Priority Threshold: 10 (generators with priority greater than this will never export)

Source: SERC Staff

For each hour of the year:

1. Update generation availability of storage facility based on amount of energy stored and discharge efficiency.

2. For each generator: a. Load available generation for this hour b. If priority < export priority: i. initialize generator as supplying all available power c. If priority > export priority: i. initialize generator as supplying the minimum turndown power.

3. For each generator in ascending order of priority: a. If demand exceeds supply and available generation exists for this generator: i. Increase generator until demand is met or full capacity is reached.

4. Test special case when import and export are simultaneously occurring. a. If both occurring:

157 i. Turn import down until export becomes zero or import becomes zero, whichever occurs first.

5. Test spinning reserve criterion: sum( gen. production * reserve fraction) >= reserve requirement for the hour. a. If reserve not met, then for each generator with reserve capability in reverse order of priority: i. Turn down generator until reserve requirement is met or minimum turndown is reached.

6. Test that turndowns haven’t reduced supply below demand: a. If demand is not being met, then for each generator that has no capacity for reserve (import will typically be the only generator that meets this criterion): i. Increase generator until demand is met or generator capacity is reached.

7. Test transmission criterion: export <= transmission constraint. a. If export exceeds transmission limit, then for each generator in reverse order of priority: i. If a regular generator (non-storage), turn down the generator until one of the following occurs: the transmission criterion is satisfied or the minimum turndown is reached. ii. If the generator is a storage facility, then the minimum turndown will actually be negative reflecting that the facility can become additional load. The minimum turndown for the facility will be set according the facilities operating strategy.

D.5 Optimization Algorithm – Differential Evolution

The authors chose Differential Evolution as the optimization algorithm. Differential Evolution is a global optimization technique, also called direct search approach, developed by Rainner Storn and Kenneth Price. As described by the inventors, the approach was designed with the following beneficial properties (Storn, 1997):

1. Ability to handle non-differentiable, nonlinear and multimodal cost functions. 2. Parallelizability to cope with computation intensive cost functions. 3. Ease of use, for example few control variables to steer the minimization. These variables should also be robust and easy to choose. 4. Good convergence properties, for example consistent convergence to the global minimum in consecutive independent trials.

The set of decision variables and their limits composes the decision space. With n decision variables, the decision space is n-dimensional. Direct search optimization approaches explore the n-dimensional decision space with a set of NP particles, each of which is an n-dimensional vector containing a value for each decision variable. In each iteration, or generation, of the algorithm, the particles are potentially updated via some scheme (multiple approaches are

158 possible). The scheme involves evaluating the objective function for a set of new candidate particles and then choosing to replace or retain the particles from the previous generation with some or all of the new particles. The process repeats until a set of convergence criteria are satisfied.

The scheme for differential evolution used in this study is described below (Storn, 1997):

1. For each of the NP particles: a. Select 3 other particles, A, B, and C, which must be distinct from each other and from the originating particle, O. b. Create candidate particle P using the following vector arithmetic: P = A + F(B-C), where F is a constant scaling parameter which lies on the range [0,2]. c. For n-1 dimensions of the candidate particle do the following: i. perform a random draw from the uniform distribution between 0 and 1. ii. test if the random value is greater than CR, another parameter of the algorithm called the cross-over parameter that lies on [0,1]. iii. if the random value exceeds the parameter, replace the value in particle P with the corresponding value in particle O. d. Evaluate the objective function, or fitness, of the candidate particle: Obj(P). e. If Obj(P) < Obj(O), then replace particle O with particle P 2. Test the stopping criteria, if not met, go to step 1.

Based on the recommendations in Pederson (2010), the dimensionality of the problem, and the computation resources at the authors’ disposal, the values in Table 41 were chosen for the differential parameters in the authors’ implementation of the optimization. Table 41: Differential Evolution Parameter Used in the Regional Energy Sector Optimization Model Parameter Description Value

n Dimensionality of the decision space 11

NP Number of Particles 69

F Scaling factor 0.66

CR Cross-over rate 0.94

Maxiter Maximum number of iterations 250

Source: SERC Staff based on Pederson (2010)

The stopping criteria test that the value of the objective function for all NP are close to each other as measure by the percent difference between each particle and the minimum value. Depending on the objective function used, the value chosen for the threshold ranged from 0.05 percent to 0.1 percent.

159

D.6 Technical and Economic Assumptions

The following tables (Table 42 through Table 49) present key technical and economic assumptions by technology type as used in the Regional Energy Planning Optimization Program (REPOP). When relevant, data sources and/or justifications are attributed to the model assumptions. Table 42: Technical and Economic Assumptions for the Natural Gas Submodel

Technology Natural Gas Power Plant Sector Electric Power Generation Source/ Units Justification Magnitude Total Resource 163 MW Existing plant; no Current Day 163 MW new generation Minimum 2030 163 MW assumed. Maximum 2030 163 MW Production Assumed that maintenance is staged and therefore Data Sources Characteristics the plant can produce any hour of the year. Capacity Factor Variable %

Fixed Costs Capital 1450 $/kW Operations 27 $/(kW*year) PG&E9 Variable Costs Operations 10 $/MWh Fuel 48 $/MWh SERC Staff10 Economic Lifetime 25 years Wartsila11 lifetime jobs/MW Impact Local Jobs 0.2, 0.07 (construction, operation) JEDI12 $/kW Economic 28, 8 (construction, Output operation) GHG Emission 0.45 tCO2e/MWh Wartsila11

Source: SERC Staff

9 PG&E Application for Certification of the Humboldt Bay Repowering Project (Energy Commission, 2007).

10 Linear extrapolation of natural gas prices. Source data from U.S. Energy Information Agency.

11 Wartsila Technical Specifications for 50DF Engine: http://www.wartsila.com/

12 National Renewable Energy Laboratory, Jobs and Economic Development Impact Natural Gas model customized to Humboldt County (SERC 2011).

160 Table 43: Technical and Economic Assumptions for the Wind Submodel

Technology Wind Sector Electric Power Generation Source/ Units Justification Magnitude Total Resource 400 MW Current Day ~0 MW SERC (2005) Minimum 2030 0 MW Maximum 2030 250 MW

Production Modeled hourly wind farm output from NREL Date Sources Characteristics Western Wind Dataset

Capacity Factor 27 %

Fixed Costs Capital 2107 $/kW Operations 24 $/(kW*year) Variable Costs Operations 0 $/MWh JEDI13 Fuel 0 $/MWh Lifetime Economic lifetime 25 years jobs/MW Impact Local Jobs 0.88, 0.15 (construction, operation) JEDI13 $/kW Economic Output 89, 23 (construction, operation)

GHG Emission 0 tCO2e/MWh

Source: SERC Staff

13 National Renewable Energy Laboratory, Jobs and Economic Development Impact model for wind default cost values for a 50MW project (SERC 2011).

161 Table 44: Technical and Economic Assumptions for the Wave Submodel

Technology Wave Sector Electric Power Generation Source/ Units Justification Magnitude Total Resource 1000 MW Current Day 0 MW SERC (2005) Minimum 2030 0 MW Maximum 2030 100 MW Production Observed Spectral Wave Data from NOAA Date Sources Characteristics Buoy14 46022 Model Capacity Factor 23 % Simulation Fixed Costs Capital 2590 $/kW KEMA Operations 36 $/(kW*year) (2009) Variable Costs Operations 12 $/MWh Fuel 0 $/MWh Lifetime Economic lifetime 25 years jobs/MW Impact Local Jobs 1.0, 9.2 (construction, operation) SERC Staff15 $/kW Economic Output 264, 831 (construction, operation) GHG Emission 0 tCO2e/MWh

Source: SERC Staff

14 National Oceanic and Atmospheric Administration, National Data Buoy Center -- Buoy 46022: http://www.ndbc.noaa.gov/station_page.php?station=46022

15 Wave Economic Impact Assessment Model (SERC 2011).

162 Table 45: Technical and Economic Assumptions for the Biomass Submodel

Technology Biomass Sector Electric Power Generation Source/ Units Justification Magnitude Total Resource 225 MW Current Day 61 MW SERC (2005) Minimum 2030 61 MW Maximum 2030 225 MW Production Observed production from Fairhaven Power Data Sources Characteristics Plant Model Capacity Factor 68 % simulation Fixed Costs Capital 3450 $/kW SERC Staff16 Operations 160 $/(kW*year) Klein (2010) Variable Costs Operations 10 $/MWh Hackett Fuel 32 $/MWh (2010) Lifetime Economic lifetime 25 years jobs/MW Impact Local Jobs 2.5, 2.0 (construction, operation) SERC Staff17 $/kW Economic Output 494, 376 (construction, operation)

GHG Emission 0.006 tCO2e/MWh Gas co-fire18

Source: SERC Staff

16 Average of various recent authoritative studies circa 2008-10 (Klein, 2010; Lazard, 2008; E3, 2008).

17 Biomass Impact Assessment Model (SERC 2011).

18 Based on production data for Fairhaven biomass power plant where natural gas co-fire is occasionally used (Marino, 2010).

163 Table 46: Technical and Economic Assumptions for the Hydropower Submodel

Technology Hydropower Sector Electric Power Generation Source/ Units Justification Oscar Magnitude Total Resource 60 MW Larson & Assoc. (1982) Current Day 10.4 MW Minimum 2030 10.4 MW SERC (2005) Maximum 2030 35 MW

Production data for Matthews Dam (HBMWD, Production Data Sources 2010), production data for Zenia Power (Burgess, Characteristics 2010), rainfall data from NOAA.

Model Capacity Factor 59 % simulation Fixed Costs Capital 4500 $/kW SERC Staff19 HBMWD Operations 124 $/(kW*year) (2010) Variable Costs Operations 0 $/MWh Fuel 0 $/MWh Lifetime Economic lifetime 40 years SERC Staff20 jobs/MW Impact Local Jobs 4.9, 1.1 (construction, operation) SERC Staff20 $/kW Economic Output 1179, 111 (construction, operation) GHG Emission 0 tCO2e/MWh

Source: SERC Staff

19 Based on data obtained from various sources (OOE, 2003; KEMA, 2009; Burgess, 2010).

20 River Hydropower Impact Assessment Model (SERC 2011).

164 Table 47: Technical and Economic Assumptions for the Photovoltaic Submodel

Technology Photovoltaic Sector Electric Power Generation Source/ Units Justification

Magnitude Total Resource 4.4 kWh/m2/day NREL (2010)

Energy Current Day 1.1 MW Commission (2010b) Minimum 2030 1.1 MW

Maximum 2030 10 MW

Production Data Sources National Solar Radiation Database (NREL, 2010). Characteristics

Model Capacity Factor 14 % simulation Fixed Costs Capital 9090 $/kW Operations 11 $/(kW*year) 21 Variable Costs Operations 0 $/MWh JEDI Fuel 0 $/MWh Lifetime Economic lifetime 20 years jobs/MW Impact Local Jobs 16, 0.1 (construction, operation) JEDI21 $/kW Economic Output 1460, 6.3 (construction, operation) GHG Emission 0 tCO2e/MWh

Source: SERC Staff

21 National Renewable Energy Laboratory, Jobs and Economic Development Impact model for Solar PV, customized for Humboldt County. Assumes 20 percent large commercial (2 100kw systems), 15 percent small commercial (6 25kw systems), 5 percent residential new construction (17 3kw systems), and 60 percent residential retrofit (200 3 kw systems). The cost data are a weighted average cost for the systems specified.

165 Table 48: Technical and Economic Assumptions for the Import/Export Submodel

Technology Import/Export Sector Electric Power Generation Source/ Units Justification Magnitude Total Resource unlimited Current Day ~60 MW PG&E (2005) Minimum 2030 60 MW Maximum 2030 250 MW Production Assumed to be available for import or export any Data Sources Characteristics hour of the year.

Model Capacity Factor Variable % simulation Fixed Costs Capital 3000 $/kW PG&E (2005) Operations 0 $/(kW*year) Variable Costs Operations 0 $/MWh Fuel Variable $/MWh SERC Staff22 Lifetime Economic lifetime 80 years

jobs/MW Data not Impact Local Jobs N/A (construction, available operation) $/kW Data not Economic Output N/A (construction, available operation) E3 GHG GHG Emission 0.36 tCO2e/MWh Tool for Buildings23 Source: SERC Staff

22 Hourly data estimated from California Independent System Operators locational marginal pricing for "Humboldt Zone" (CAISO, 2010).

23 Energy, Environment, and Economics Greenhouse Gas Tool for Buildings: http://www.ethree.com

166 Table 49: Technical and Economic Assumptions for the Pumped Hydro Storage Submodel

Technology Pumped Hydro Storage Sector Electric Power Generation Source/ Units Justification

Magnitude Total Resource Undetermined Current Day 0 MW Minimum 2030 0 MW Maximum 2030 25 MW Assumed to be available for discharge/charge Production Data Sources any hour of the year given that stored energy is Characteristics available/vacant respectively. Round trip Peterson 78 % efficiency (2008) Fixed Costs Capital 1250 $/kW Operations 65 $/(kW*year) Variable Costs Operations 20 $/MWh Klein (2008) Fuel 0 $/MWh Lifetime Economic lifetime 30 years jobs/MW Data not Impact Local Jobs N/A (construction, available operation) $/kW Data not Economic Output N/A (construction, available operation) GHG Emission 0 tCO2e/MWh

Source: SERC Staff

167 D.7 References

Auffhammer, Maximilian and Anin Aroonruengsawat (2010). “Uncertainty over Population, Prices or Climate? Identifying the Drivers of California’s Future Residential Electricity Demand.” Energy Institute at Haas Working Paper Series. August, 2010.

Burgess, Ross (2010), owner of Zenia micro-hydropower plant. Personal communication.

California Department of Finance (2007). Population Projections for California and Its Counties 2000-2050, by Age, Gender and Race/Ethnicity. Sacramento, California, July 2007.California Energy Commission (2008) 2008 DEER Update – Summary of Measure Energy Analysis Revisions. Available online: http://www.deeresources.com [Accessed August 12, 2010].California Energy Commission (2010a). 2004-2008 Humboldt County Electricity and Natural Gas Billed Usage Data by Sector. Personal Communication via email, 02/2010

California Energy Commission (2010b). California Solar Photovoltaic Statistics & Data. http://energyalmanac.ca.gov/renewables/solar/1998-2006.html.California Public Utilities Commission (2010), CPUC Approves PG&E request to match U.S. Department of Energy award for compressed air energy storage project. Press release Docket #: A.09-09-019

Chi-Jen, Y.; Williams, E. (2009). “Energy Storage for Low-Carbon Electricity.” Duke University Climate Change Policy Partnership. http://www.nicholas.duke.edu/ccpp/ccpp_pdfs/energy.storage.pdf

Deane, J.P., Gallachoir, B.P. O, & McKeogh, E.J., (2010). Techno-economic review of existing and new pumped hydro energy storage plant. Renewable and Sustainable Energy Review 14(2010) 1293-1302

Electric Power Research Institute (2003). EPRI-DOE Handbook of Energy Storage for Transmission and Distribution Applications. Palo Alto, CA, and the U.S. Department of Energy, Washington, DC: 2003. 1001834.

Electricity Storage Association. Accessed December 2010 at http://www.electricitystorage.org/site/home/

Energy Information Agency, Annual Energy Outlook 2011, Energy Generating Capacity http://www.eia.gov/analysis/projection-data.cfm#annualproj [Accessed April 7, 2011]

Enshen, Long (2005). Research on the influence of air humidity on the annual heating or cooling energy consumption, Building and Environment, Volume 40, Issue 4, April 2005, p. 571-578.

Hackett, Steven (2010). Personal communication with a Humboldt-based biomass power plant operator. Humboldt State University, Department of Economics.

Heller, A.J. (2002). Heat-load modelling for large systems. Applied Energy 72 (2002) p. 371-387.

Humboldt Bay Municipal Water District (2010). Production data for hydropower plant at Matthew’s Dam, Ruth Lake.

168 Iowa Stored Energy Park Project (2011). Economics Study Summary. http://www.isepa.com/index.asp [Accessed April 7, 2011].

Itron (2005). 2004-2005 Database for Energy Efficiency Resources (DEER) Update Study: Final Report. Available online: http://www.deeresources.com [Accessed August 12, 2010].

KEMA Inc. (2009). Renewable Energy Cost of Generation Update. California Energy Commission, CEC-500-2009-084.

Klein, R. (2008). "Symbiotics LLC. Presentation to NWPCC" (presentation to the Northwest Power and Conservation Council, Portland Oregon, October 17, 2008)

Klein, Joel. (2010). Comparative Costs of California Central Station Electricity Generation Technologies, California Energy Commission, CEC-200-2009-017-SD.

Marino, Bob (2010), Power plant production for DG Fairhaven Power LLC. Furnished personally by Bob Marino, General Manager.

National Geographic News (2010) “Texas Pioneers Energy Storage in Giant Battery.” March 25, 2010. http://news.nationalgeographic.com/news/2010/03/100325-presidio-texas-battery/ [Accessed December 2010].

National Renewable Energy Laboratory (2010). National Solar Radiation Database: 1991-2005 Update. Arcata Airport – Site 725945.

Oregon Office of Energy (2003). Case Study: Micro-hydro Project – History repeats itself at Crown Hill Farm. Oregon Office of Energy: OOE 11/02 CF-074, January, 2003.

Oscar Larson & Associates (1982). An Analysis of Small Hydroelectric Planning Strategies. Prepared for: Humboldt County Board of Supervisors, May, 1982.

Pacific Gas & Electric Company (2005). Humboldt Long Term Transmission Assessment, Draft Study Report. Prepared by: Dave Casuncad, Electric T&D Engineering. Prepared for: Pacific Gas and Electric Company, January 13, 2005.

Pacific Gas & Electric (2010a), WaveConnect Humboldt Working Group Communication, November 2010.

Pacific Gas & Electric (2010b) Cogeneration and Small Power Production Semi-Annual Report, July 2010: http://www.pge.com/includes/docs/pdfs/b2b/qualifyingfacilities/cogeneration/jul2010coge n.pdf

Pacific Gas & Electric (2010c), Data furnished to Schatz Energy Research Center by Pacific Gas & Electric for the Humboldt County Renewable Energy Secure Community study.

Pederson, Magus Erik Hvass (2010). “Good Parameters for Differential Evolution.” Hvass Laboratories, Technical Report No. HL1002, 2010.

169 RECS (2009). Public Microdata: Residential Buildings Energy Consumption Survey, 2005. Energy Information Administration. http://www.eia.doe.gov/emeu/recs/recspubuse05/pubuse05.html [Accessed July 12, 2010]

Schatz Energy Research Center (2005). Humboldt County Energy Element Appendices: Technical Report, Prepared for Redwood Coast Energy Authority.

Schatz Energy Research Center (2011). Humboldt County as a Renewable Energy Secure Community: Economic Analysis Report. Public Interest Energy Research Program Interim Project Report. Prepared for: California Energy Commission, April 2011.

Storn, Rainer and Kenneth Price (1997). “Differential Evolution – A Simple and Efficient Hueristic for Global Optimization over Continuous Spaces.” Journal of Global Optimization 11:341-359, 1997.

Succar, S., & Willams, R. H. (2008). Compressed Air Energy Storage: Theory, Resources, and Applications for Wind Power

USA Today. July 5, 2007. “New battery packs powerful punch.” http://www.usatoday.com/money/industries/energy/2007-07-04-sodium-battery_N.htm [Accessed December 2010]

U.S. Census (2000). United States Census 1990 and 2000 Results available online http://factfinder.census.gov [Accessed August 16, 2010].

Wojdyga, Krzysztof (2008). An influence of weather conditions on heat demand in district heating systems. Energy and Buildings 40 (2008) p. 2009-2014.

3Tier (2010), “Development of Regional Wind Resource and Wind Plant Output Datasets.” Subcontract Report, NREL/SR-550-47676. National Renewable Energy Laboratory, U.S. Department of Energy. March 2010.

170 ATTACHMENT I: PG&E Interconnection Feasibility Study Plan

(This page intentionally left blank)

171 Interconnection Feasibility Study Plan

Generation Interconnection

Humboldt County Renewables Project

Pacific Gas and Electric Company

WE DELIVER ENERGY.

February 11, 2011

172

Table of Contents

1. Introduction...... 1

2. Schedule...... 2

3. Cost Estimates ...... 2

4. Project and Interconnection Information...... 2

5. Study Assumptions ...... 3

6. Power Flow Study Base Cases ...... 4

7. Study Scope...... 6

7.1 Steady State Power Flow Analysis ...... 6

7.2 Reactive Power Deficiency Analysis...... 7

8. Technical Requirements ...... 7

173 INTERCONNECTION FEASIBILITY STUDY PLAN HUMBOLDT COUNTY RENEWABLES PROJECT FEBRUARY 1, 2011

1. Introduction

Pacific Gas and Electric Company (PG&E) is participating in the Redwood Coast Energy Authority's (RCEA) Exploratory Project under the California Energy Commission (CEC) Renewable-based Energy Secure Communities (RESCO) Grant Program.

The Schatz Energy Research Center of the Humboldt State University requested that PG&E complete an Interconnection Feasibility Study (FS) to evaluate the transmission impacts of the proposed Humboldt County Renewables Project (Project) assuming a projected year 2030 peak electric loading condition. Scenario 1 of the Project consists of nine (9) proposed new generation facilities in Humboldt County with a total generation output of 253 MW.

Table 1-1: Scenario 1 New Generation Projects

Max Location Technology MW Description

Wind 50 Bear River Ridge Wind 75 Bear River Ridge Biomass 40 Samoa Biomass 15 Willow Creek Biomass 15 Garberville Hydro 8 Maple Creek Hydro 5 Hoopa Valley Storage – Pumped Hydro 15 Ruth Lake Wave 30 Samoa Total New Generation 253

The FS will identify:

x Transmission system impacts caused solely by the addition of the Project;

x System reinforcements necessary to mitigate any adverse impacts of the Project under various system conditions; and

x Facilities required for system reinforcements with a non-binding good faith estimate of cost responsibility.

The Schatz Energy Research Center requested to perform the peak studies using a Humboldt area load of 253 MW, which is representative of the year 2030 winter peak. Typically, as required by the North American Reliability Corporation (NERC) reliability standards, PG&E’s transmission planning studies focus on a 10 year planning horizon for long term studies to only identify marginal conditions that may have longer lead-time solutions. Therefore, the FS will only be a preliminary evaluation of the feasibility and the magnitude of the system impacts resulting from the proposed generation interconnection with the PG&E transmission system.

1

174 INTERCONNECTION FEASIBILITY STUDY PLAN HUMBOLDT COUNTY RENEWABLES PROJECT FEBRUARY 1, 2011

2. Schedule

Table 2-1 shows the milestones/schedules associated with the study.

Table 2-1: Study Schedule Target Task Milestone Description Date 1 PG&E issues Interconnection Feasibility Study Plan February 14, 2011 2 PG&E issues results to Humboldt County March 2, 2011

3. Cost Estimates

The FS will provide a list of required facilities with a non-binding good faith estimate of cost responsibility of the facilities necessary to interconnect the Project. These costs have no associated degree of accuracy and are provided for informational purpose only.

4. Project and Interconnection Information

Scenario 1 of the Project includes nine new generation facilities. The facilities and the proposed interconnection points are summarized in Table 4-1.

Table 4-1: Scenario 1 New Generation Projects

Max Location Proposed Point of Technology MW Description Interconnection

Wind 50 Bear River Ridge Rio Dell Substation 60 kV Bus Wind 75 Bear River Ridge Rio Dell Substation 60 kV Bus Biomass 40 Samoa Fairhaven Substation 60 kV Bus Biomass 15 Willow Creek Willow Creek Substation 60 kV Bus Biomass 15 Garberville Garberville Substation 60 kV Bus Hydro 8 Maple Creek Maple Creek Substation 60 kV Bus Hydro 5 Hoopa Valley Hoopa Substation 60 kV Bus Storage – Pumped Hydro 15 Ruth Lake Low Gap Substation 115 kV Bus Wave 30 Samoa Fairhaven Substation 60 kV Bus

Figure 5-1 provides the map for the Project and the transmission facilities in the vicinity.

2

175 INTERCONNECTION FEASIBILITY STUDY PLAN HUMBOLDT COUNTY RENEWABLES PROJECT FEBRUARY 1, 2011

Figure 5-1: Vicinity Map

5. Study Assumptions

PG&E will conduct the FS under the following assumptions:

1) The total new generation output of the Project to the PG&E transmission system is 253 MW. New generation facilities will be modeled as shown in Table 5-1 with the full MW output at each location:

3

176 INTERCONNECTION FEASIBILITY STUDY PLAN HUMBOLDT COUNTY RENEWABLES PROJECT FEBRUARY 1, 2011 Table 5-1: Scenario 1 New Generation Projects

Max Proposed Point of Interconnection MW

50 Rio Dell Substation 60 kV Bus 75 Rio Dell Substation 60 kV Bus 40 Fairhaven Substation 60 kV Bus 15 Willow Creek Substation 60 kV Bus 15 Garberville Substation 60 kV Bus 8 Maple Creek Substation 60 kV Bus 5 Hoopa Substation 60 kV Bus 15 Low Gap Substation 115 kV Bus 30 Fairhaven Substation 60 kV Bus

2) This study will take into account the planned generating facilities in PG&E’s service territory that are in the California Independent System Operator Corporation (CAISO) Generation Interconnection Queue in the vicinity of the Project. In addition, all CAISO approved PG&E transmission projects that will be operational by 2020 will also be included.

3) As requested by the Schatz Energy Research Center, the load Humboldt area load to be studied is 253 MW and it represents the winter peak loading condition for the year 2030. The study will utilize a 2020 Winter Peak Base Case incorporating the year 2030 load as requested.

Note: Based on current observed lower load trends, the Humboldt area load may not reach the projected levels being assumed for the year 2030.

6. Power Flow Study Base Cases

Power flow analyses will be performed to ensure that PG&E’s transmission system remains in full compliance with North American Reliability Corporation (NERC) reliability standards TPL-001, 002, 003 and 004 with the proposed interconnection. The results to these power flow analyses will serve as documentation that an evaluation of the reliability impact of new generation facilities and their connections on interconnected transmission systems is performed. If a NERC reliability deficiency exists as a result of this new generation interconnection, PG&E will identify the problem and develop appropriate transmission solutions to comply with NERC reliability standards.

Two (2) power flow base cases will be used to evaluate the transmission system impacts of this interconnection. While it is impractical to study all combinations of system load and generation levels during all seasons and at all times of the day, these base cases represent extreme loading and generation conditions for the study area.

PG&E cannot guarantee that the Project:

4

177 INTERCONNECTION FEASIBILITY STUDY PLAN HUMBOLDT COUNTY RENEWABLES PROJECT FEBRUARY 1, 2011 a) can operate at maximum rated output 24 hours a day, year round, without system impacts; nor

b) will not have system impacts during the times and seasons not studied in the FS.

The following power flow base cases will be used for the analysis and FS:

„ 2020 Winter Peak Base Case:

Power flow analysis will be performed using PG&E’s 2020 Winter Peak Base Case (in General Electric Power Flow format). This base case was developed from PG&E’s 2010 base case series and has a 1-in-10 year adverse weather load level for PG&E’s Coastal areas.

„ 2020 Off-Peak Base Case:

Power flow analysis will also be performed using PG&E’s 2020 Off-Peak Base Case (in General Electric Power Flow format) to evaluate potential congestion on transmission facilities during the Off-Peak system conditions.

These base cases will model all CAISO approved PG&E transmission projects that would be operational by 2020. The planned PG&E system additions and upgrades are summarized in Table 6-1.

Table 6-1: Planned PG&E System Additions and Upgrades

Project Fulton - Fitch Mountain 60 kV Line Reconductoring Garberville Reactive Support Humboldt - Harris 60 kV Reconductoring Humboldt 115/60 kV Transformer Replacements Humboldt Reactive Support Maple Creek Reactive Support Mendocino Coast Reactive Support

The base cases will also include all proposed generation projects that in the CAISO Generation Interconnection Queue in the vicinity of the Project and their associated identified network upgrades. However, some generation projects that are electrically far from the Project will be either turned off or modeled with reduced generation to balance the loads and resources in the power flow model. The major generation projects included and the associated network upgrades are summarized below in Table 6-2 and Table 6-3, respectively.

5

178 INTERCONNECTION FEASIBILITY STUDY PLAN HUMBOLDT COUNTY RENEWABLES PROJECT FEBRUARY 1, 2011 Table 6-2: Proposed Humboldt Area Generation Interconnection Projects

CAISO Max Commercial Point of Interconnection Fuel Queue MW Operation Date

38 Humboldt Power Plant Substation Natural Gas 146.4 9/30/2010 212 Rio Dell Substation 60 kV bus Wind 50 12/15/2013 472 Ultra Power 60 kV Tap Line Biomass 13.8 4/23/2010 Total 210.2

Table 6-3: Network Upgrades for Proposed Generation Projects

Project Reconductor Humboldt Bay - Rio Dell Jct 60 kV Line Reconductor Rio Dell 60 kV Tap Line Reconductor Humboldt Bay - Eureka 60 kV Line Reconductor Humboldt - Humboldt Bay #2 60 kV Line

7. Study Scope

The FS will determine the impact of the Project on the PG&E transmission system. The specific studies to be conducted are outlined below:

7.1 Steady State Power Flow Analysis

The CAISO controlled-grid Reliability Criteria, which incorporates the Western Electricity Coordinating Council (WECC) and NERC planning criteria, will be used to evaluate the impact of the Project on the PG&E transmission system.

Power Flow analysis will be performed using the base cases described in Section 6. These base cases will be used to simulate the impact of the Project during normal (CAISO Category “A”) operating conditions as well as during single (CAISO Category “B”) and selected multiple (CAISO Category “C”) contingency conditions. The study will cover the transmission facilities within PG&E’s Humboldt and North Coast Areas.

The types of contingencies to be evaluated under each category are summarized below Table 7-1.

6

179 INTERCONNECTION FEASIBILITY STUDY PLAN HUMBOLDT COUNTY RENEWABLES PROJECT FEBRUARY 1, 2011 Table 7-1: Summary of Planning Standards Contingencies Description CAISO Category “A” All facilities in service – Normal Conditions x B1 - All single generator outages. x B2 - All single transmission circuit outages. CAISO Category “B” x B3 - All single transformer outages. x Selected overlapping single generator and transmission circuit outages for the transmission lines and generators. x C1 - SLG Fault, with Normal Clearing: Bus outages (60-230 kV) x C2 - SLG Fault, with Normal Clearing: Breaker failures (excluding bus tie and sectionalizing breakers) at the same bus section above. x C3 - Combination of any two-generator/transmission line/transformer outages. CAISO Category “C” x C4 - Bipolar (dc) Line x C5 - Outages of double circuit tower lines (60-230 kV) x C6 - SLG Fault, with Delayed Clearing: Generator x C7 - SLG Fault, with Delayed Clearing: Transmission Line x C8 - SLG Fault, with Delayed Clearing: Transformer x C9 - SLG Fault, with Delayed Clearing: Bus Section

Although most of the CAISO Category “C” contingencies will be considered for this study, it is impractical to evaluate all the CAISO Category “C” contingencies. For this reason, select critical Category “C” contingencies (C1 – C9) will be evaluated as part of this study.

7.2 Reactive Power Deficiency Analysis

With the generation project included in the system model, CAISO Category “B” and “C” contingencies will be analyzed to identify any reactive power deficiency:

x Whether the results show voltage drops of 5% or more from the pre- Project levels, or

x Whether the results fail to meet applicable voltage criteria.

A post-transient power flow analysis will be performed, if deemed necessary, after considering the network topology or power transfer paths involved when a significant amount of power transfer occurs.

8. Technical Requirements

The PG&E Interconnection Handbook explains the technical requirements for interconnection of loads and generators to PG&E’s transmission system. The Interconnection Handbook documents facility connection requirements to the PG&E system as required in NERC Standard FAC-001-0. They are based on applicable FERC and CPUC rules and tariffs (e.g., Electric Rules 2, 21 and 22), as well as accepted industry practices and standards. In addition to providing reliability, these technical requirements are consistent with safety for PG&E workers and the public.

7

180 INTERCONNECTION FEASIBILITY STUDY PLAN HUMBOLDT COUNTY RENEWABLES PROJECT FEBRUARY 1, 2011 The PG&E Interconnection Handbook applies to Retail and Wholesale Entities, which own or operate generation, transmission, and end user facilities that are physically connected to, or desire to physically connect to PG&E’s electric system. All technical requirements described or referred to in the Handbook apply to new or re-commissioned Generation Facilities. The Generation Interconnection Handbook comprising sections G-1 through G-5 applies to Generation Entities.

PG&E has established standard operating, metering and equipment protection requirements for loads and generators. The Interconnection Handbook covers such requirements for all transmission-level load and generation entities wishing to interconnect with PG&E’s electric system. Additional, project-specific requirements may apply.

The PG&E Interconnection Handbook includes, but is not limited to such operating requirements as the following:

x The Project must be able to meet the power factor requirements of 90 percent lagging and 95 percent leading.

x The Project must have Automatic Voltage Regulation (AVR) and be able to maintain the generator voltage under steady-state conditions within ±0.5 percent of any voltage level between 95 percent and 105 percent of the rated generator voltage.

Generators must also meet all applicable CAISO, NERC, and WECC standards. NERC and WECC standards include, but are not limited to such requirements as the following:

x The Project must be able to remain on line during voltage disturbances up to the time periods and associated voltage levels as required by the WECC Low Voltage Ride Through (LVRT) standards that are in-line with FERC Order No. 161-A. The WECC LVRT standard is available on the WECC web site at:

http://www.wecc.biz/committees/StandingCommittees/PCC/TSS/Shared %20Documents/Voltage%20Ride%20Through%20White%20Paper.pdf

x Currently NERC is working on a Voltage Ride Through standard, PRC-024-1, that would be applicable to all generators interconnecting to the transmission grid. Until PRC-024-1 is effective, PG&E and the CAISO will require that all generators comply with the existing WECC LVRT requirements. The PRC- 024-1 standard Draft 1 can be found on the NERC web site at

http://www.nerc.com/docs/standards/sar/PRC-024- 1_Draft1_2009Feb17.pdf

All generators must satisfy the requirements of the PG&E’s Interconnection Handbook and meet all applicable CAISO, NERC, and WECC standards. PG&E will not agree to interconnect any new generators unless all technical and contractual requirements are met.

8

181 INTERCONNECTION FEASIBILITY STUDY PLAN HUMBOLDT COUNTY RENEWABLES PROJECT FEBRUARY 1, 2011 The interconnecting customer should be aware that the information in the PG&E Interconnection Handbook is subject to change. Parties interconnecting to the PG&E electric system should verify with their PG&E representative that they have the latest versions. The PG&E Interconnection Handbook is available on the PG&E web site at:

http://www.pge.com/about/rates/tariffbook/ferc/tih/

9

182 INTERCONNECTION FEASIBILITY STUDY PLAN HUMBOLDT COUNTY RENEWABLES PROJECT FEBRUARY 1, 2011

Feasibility Study Agreement

The Schatz Energy Research Center of the Humboldt State University has reviewed the Feasibility Study Plan for interconnection of the Humboldt County Renewables Project to PG&E's transmission system in Humboldt County, State of California, and agrees with the proposed plan.

Dated this day of February, 2011

Schatz Energy Research Center of the Humboldt State University:

BY: ______(Signature)

______(Type or Print Name)

MAILING ADDRESS:

______

______

______

______

1

183 ATTACHMENT II: PG&E Interconnection Feasibility Study Results

(This page intentionally left blank)

184 Humboldt County RESCO - Interconnection Feasibility Study

Generation Interconnection

Additional Information Request

Pacific Gas and Electric Company

WE DELIVER ENERGY.

June 10, 2011

185 Pacific Gas and Electric Company (PG&E) is participating in the Redwood Coast Energy Authority's (RCEA) Exploratory Project under the California Energy Commission (CEC) Renewable-based Energy Secure Communities (RESCO) Grant Program.

In March 2011, PG&E completed an Interconnection Feasibility Study (FS) to evaluate the transmission impacts of the proposed Humboldt County Renewables Project (Project) assuming a projected year 2030 peak electric loading condition.

Following completion of the FS, the Schatz Energy Research Center of the Humboldt State University requested that PG&E provide additional support by responding to the following questions:

1. Provide background narrative on the Humboldt County transmission system. Identify weak points (e.g., substations) and structural weaknesses (e.g., 60 kV as the main system voltage) and discuss the implications for shifting from being a load pocket to a generation pocket. 2. Provide first order cost estimates for system upgrades that would enable RESCO generators to meet NERC requirements. 3. Identify key bottlenecks in the system and provide recommendations on key upgrades that would be most cost-effective in enabling the build-out of future renewable generation. Also, identify geographic areas that are particularly problematic for adding new generation. 4. Explain whether or not the various system upgrades would be considered “Network Upgrades” according to CAISO and whether or not the interconnection customer would be repaid the cost of associated Network Upgrades over a five-year period commencing on the Commercial Operation Date of their newly installed generation.

Responses to the above questions and supporting information are provided in this document.

2

186 1. Provide background narrative on the Humboldt County transmission system. Identify weak points (e.g., substations) and structural weaknesses (e.g., 60 kV as the main system voltage) and discuss the implications for shifting from being a load pocket to a generation pocket.

Background Electric customers in the Humboldt Area are served by the Pacific Gas and Electric Company’s network of 115 kV and 60 kV transmission lines. The Humboldt electrical system receives its power primarily from the PG&E’s Humboldt Generating Station and is connected to Mendocino and Cottonwood sources through the Bridgeville-Garberville and Garberville-Laytonville 60 kV lines respectively and Bridgeville – Cottonwood and Trinity – Cottonwood 115 kV Lines. The 115 kV lines from Cottonwood are over 100 miles long and the 60 kV lines from Trinity and Mendocino are 55 and 80 miles long, respectively. Electric customers within the Humboldt area are mainly served by the Company’s network of 60 kV transmission lines and distribution substations

The Humboldt transmission system is currently operated as a balanced system of internal generation with some imports and has not been designed to withstand exporting large amounts of power outside of its system. The system has one major 115 kV loop to transport power throughout the Humboldt system while the load is served primarily through the 60 kV transmission system. Generally, the 60 kV transmission system is networked together and is comprised of long circuits with enough capacity to serve the local area load. Consequently, as the power flow analysis with the proposed generation shows, the 60 kV transmission system is not adequate to transport mid to large amounts of generation and would be considered the main limitation or system weakness in converting Humboldt into a generation exporting area. Excess generation in this system results in many normal overloads of the 60 kV system and during contingencies, the same amount of generation needs to be pushed through the rest of the 60 kV circuits causing even more severe overloads. Ideally, it would make sense to add new generation close to a major transmission line in order to transport the electric power, however if this preferred route was compromised for any reason the rest of the networked 60 kV system would have to pick up the flow. This again would lead to major overloads throughout the entire Humboldt system. Another major limitation for interconnecting large amounts of generation in Humboldt is the fact that the resulting extremely high flows of electric power on the 60 kV transmission system cause heavy reactive power losses on the lines due to the surge impedance line characteristics. This issue is of concern because it may lead to poor system reliability and possible voltage instability.

Power Flow Analyses – Identifying System Constrains The power flow analysis completed evaluated the transmission impacts of adding 9 renewable generating facilities (totaling 253 MW) in Humboldt County by the year 2030. These generating facilities would mainly be interconnected to the 60 kV system in the Humboldt area as indicated in the table below. A geographic map indicating the actual locations is also presented.

3

187

Table 1 – Proposed Generation Technology Max MW Location Proposed Point of Interconnection Description Wind 50 Bear River Ridge Rio Dell Substation 60 kV Bus Wind 75 Bear River Ridge Rio Dell Substation 60 kV Bus Biomass 40 Samoa Fairhaven Substation 60 kV Bus Biomass 15 Willow Creek Willow Creek Substation 60 kV Bus Biomass 15 Garberville Garberville Substation 60 kV Bus Hydro 8 Maple Creek Maple Creek Substation 60 kV Bus Hydro 5 Hoopa Valley Hoopa Substation 60 kV Bus Storage – Pumped Hydro 15 Ruth Lake Low Gap Substation 115 kV Bus Wave 30 Samoa Fairhaven Substation 60 kV Bus

Figure 1 – Geographic Map with Proposed Generation

The study took into account all existing and planned generating facilities in PG&E’s service territory in the California Independent System Operator Corporation (CAISO) Generation Interconnection Queue within the vicinity of the Project along with identified associated network upgrades. In total, the generation assumed in place in Humboldt before adding the proposed 9 projects is roughly 280 MW. When adding the proposed generating projects, the Humboldt area represented a total of 533 MW of generating capacity.

4

188 In addition for this study, all CAISO approved PG&E transmission projects operational by 2020 were also included and the Humboldt area load was adjusted to 253 MW to represent the winter peak loading condition for the year 2030.

Simply reviewing the load vs. generation amounts assumed in the study, the Humboldt area would be exporting 280 MW of electricity onto the rest of the PG&E system. As expected, given the limitation of the transmission system, numerous thermal violations were observed during normal system conditions and under different contingency events. Below is a summary of the facility overloads identified and that PG&E developed potential solutions for.

Category A – Thermal Problems 1. Bridgeville 115/60 kV Transformer Bank 1 2. Bridgeville-Garberville 60 kV Line (Bridgeville-Fruitland Jct) 3. Rio Dell 60 kV Tap Line (Rio Dell-Scotia Tap) 4. Rio Dell 60 kV Tap Line Scotia Tap-Rio Dell Jct 5. Rio Dell Jct-Bridgeville 60 kV Line (Rio Dell Jct-Carlotta) 6. Rio Dell Jct-Bridgeville 60 kV Line (Carlotta-Swains Flat) 7. Rio Dell Jct-Bridgeville 60 kV Line (Swains Flat-Bridgeville)

Category B – Thermal Problems 8. Bridgeville-Cottonwood 115 kV Line (Low Gap-Bridgeville) 9. Bridgeville-Cottonwood 115 kV Line (Wildwood-Cottonwood) 10. Bridgeville-Garberville 60 kV Line (Fruitland Jct-Fort Seward Jct) 11. Bridgeville-Garberville 60 kV Line (Fort Seward Jct-Garberville) 12. Essex Jct-Arcata-Fairhaven 60 kV Line (Arcata Jct 2-Fairhaven) 13. Fairhaven-Humboldt 60 kV Line (Arcata Jct 2-Humboldt) 14. Fairhaven-Humboldt 60 kV Line (Arcata Jct 2-Sierra Pacific) 15. Fairhaven-Humboldt 60 kV Line (Sierra Pacific-Fairhaven) 16. Humboldt Bay-Humboldt #1 60 kV Line (Humboldt-Humboldt Jct) 17. Humboldt Bay-Rio Dell Jct 60 kV Line (Eel River-Newburg) 18. Humboldt Bay-Rio Dell Jct 60 kV Line (Humboldt Bay-Eel River) 19. Humboldt Bay-Rio Dell Jct 60 kV Line (Newburg-Rio Dell Jct) 20. Humboldt-Bridgeville 115 kV Line 21. Humboldt-Trinity 115 kV Line 22. Humbolt Bay-Eureka 60 kV Line 23. Trinity-Cottonwood 115 kV Line (Jessup Tap-Cottonwood)

Category C – Thermal Problems 24. Trinity-Maple Creek 60 kV Line (Big Bar-Hyampom Jct) 25. Trinity-Maple Creek 60 kV Line (Maple Creek-Ridge Cabin) 26. Trinity-Maple Creek 60 kV Line (Ridge Cabin-Hyampom Jct)

Note that the power flow analysis performed did not consider the potential issues at the California-Oregon-Intertie (COI). The injection of 280 MW from the Humboldt area could potentially cause reliability and/or congestion issues that may limit the amount of generation that can be interconnected in the Humboldt area or at least affect the amount of hours when the full

5

189

capacity can be accepted into the system without first upgrading COI. These potential issues would be identified when a more detailed engineering study is performed as part of a formal interconnection request.

2. Provide first order cost estimates for system upgrades that would enable RESCO generators to meet NERC requirements.

As a registered Transmission Planner (TP), PG&E must ensure that its transmission system remains in full compliance with the North American Reliability Corporation (NERC) reliability standards TPL-001, 002, 003 and 004 at all times. To supplement the FS, PG&E developed transmission reinforcement alternatives that address all identified NERC reliability deficiencies in the Humboldt area electric transmission system resulting from the new generation interconnection. The costs estimates developed for the identified system reinforcements are based on per unit cost estimates and are non-binding.

Three alternatives were developed and studied to ensure that the identified transmission system issues are corrected. The first alternative consists of simply upgrading the facilities identified to have thermal or voltage problems, by replacing conductor and substation equipment such as transformers with higher capacity equipment and adding voltage support equipment. The second alternative involves a combination of facilities upgrades and conversion of operating voltage of certain parts of the Humboldt 60 kV system to 115 kV and 230 kV voltage operation. The third alternative considered entails building new 230 kV lines and associated facilities along with local system reinforcement as necessary. Analysis of the cost and system impacts, conclude that the third alternative, new 230 kV lines, is preferred as it provides the higher level of reliability and system integrity to the Humboldt transmission system at the lowest cost. The three alternatives are further discussed below.

Alternative 1 –Replacement of Overloaded Facilities The first alternative entails reconductoring of the majority of the lines within the Humboldt transmission system as well as the existing tie lines to Mendocino and Cottonwood. A total of 484 miles of line are proposed to be reconductored. However, simple reconductoring does not mitigate the voltage violations caused by heavy reactive power loses in the lines due to the high electric power flow on low voltage lines. To address the high reactive power loses issue, this alternative includes installation of a large amount of voltage support across the Humboldt transmission system to meet voltage criteria. If this alternative were to be selected, further reactive and voltage stability studies would need to be performed to ensure local transmission grid integrity.

This alternative is estimated to cost about $940 Million. This alternative is not recommended as it very costly and does not improve system reliability, in fact, it can make the system more vulnerable to operate.

Alternative 2 – Partial Voltage Conversion and Reconductor This alternative attempts to utilize and maximize the existing infrastructure by converting portions of the existing transmission system to higher voltage operation. The alternative entails

6

190 converting the entire southern portion of the Humboldt system from 60 kV to 115 kV operation as well as converting the inner 60 kV loops that transfer power throughout the Humboldt area to 115 kV operation. The existing 115 kV transmission lines between Humboldt and Cottonwood Substations are also converted to 230 kV operation to withstand the large amount of generation being exported.

This alternative requires a total of 427 miles of transmission line conversion to higher voltage, 92 miles of reconductoring, 6 transformer installations or replacement, voltage conversion of 25 substations and building 3 new substations. This alternative is estimated cost about $1 Billion.

Alternative 3 – New 230 kV Facilities This alternative consists of building a new 230 kV Substation at Rio Dell Jct and building a 95- mile 230 kV double circuit tower line from Rio Dell Jct to Cottonwood 230 kV Substation. Rio Dell Jct was chosen because the largest generation is proposed at this location. The line terminates at Cottonwood Substation because that is the closest 230 kV Substation to the Humboldt area. .

This alternative considers the 253 MW of proposed renewable generation at the given locations. If the amounts and/or location for the generation changes, this alternative would need to be expanded or revised to accommodate such changes. This alternative is preferred in comparison to the other alternatives studied because it allows generation in Humboldt to be connected without compromising the existing transmission system. In addition, this alternative provides the most effective way to address identified thermal and voltage violations while requiring the least amount of overall upgrades to the existing system to obtain reliability and maintain system integrity. This alternative is estimated to cost about $260 Million.

A detailed list of required facility upgrades along with non-binding cost estimates for each alternative are provided in the following tables. The costs provided are for informational purposes only. These were derived using typical per unit costs and have no associated degree of accuracy.

Furthermore and as indicated before, it is important to note that the power flow study performed and the alternatives and costs outlined above, do not account or consider limitations at the California-Oregon-Intertie (COI) that may exist now or at the time when this proposed renewable generation may develop. The injection of roughly 280 MW from the Humboldt area to the Cottonwood transmission system could potentially cause reliability or congestion issues that may limit the amount of generation that can be interconnected in the Humboldt area without any upgrades to the COI. Potential COI upgrades are unknown at this time and are not included in the costs provided. If any upgrades are required, these would be identified by detailed engineering studies when a formal interconnection request is submitted.

7

191 Table 2 – Alternative 1 Associated Work Alternative 1 - Reconductoring, Voltage Support and Transformer Upgrades only P.U. Total Description Units Costs Costs Reconductor Bridgeville - Cottonwood 115 kV Line 86.4 1.5 129.6 Reconductor Bridgeville - Garberville 60 kV Line 36.1 1.5 54.1 Reconductor Rio Dell 60 kV Tap Line 5.8 1.5 8.7 Reconductor Rio Dell Jct - Bridgeville 60 kV Line 21.3 1.5 31.9 Reconductor Essex Jct - Arcata – Fairhaven 60 kV Line 16.1 1.5 24.1 Reconductor Arcata - Humboldt 60 kV Line 7.2 1.5 10.9 Reconductor Fairhaven - Humboldt 60 kV Line 15.3 1.5 22.9 Reconductor Humboldt Bay - Humboldt No.1 60 kV Line 8.1 1.5 12.1 Reconductor Humboldt Bay - Rio Dell Jct 60 kV Line 4.4 1.5 6.6 Reconductor Humboldt - Bridgeville 115 kV Line 30.3 1.5 45.4 Reconductor Humboldt - Trinity 115 kV Line 68.7 1.5 103.0 Reconductor Humboldt Bay - Eureka 60 kV Line 5.6 1.5 8.5 Reconductor Trinity - Cottonwood 115 kV Line 45.8 1.5 68.7 Reconductor Garberville - Laytonville 60 kV Line 40.0 1.5 60.0 Reconductor Humboldt - Eureka 60 kV Line 1.1 1.5 1.7 Reconductor Humboldt - Maple Creek 60 kV Line 13.9 1.5 20.9 Reconductor Trinity - Maple Creek 60 kV Line 61.1 1.5 91.7 Reconductor Trinity - Keswick 60 kV Line 30.5 1.5 45.7 Reconductor Cascade - Keswick 60 kV Line 7.2 1.5 10.8 Install New Bridgeville No.2 115/60 kV XFMR with higher ratings 1 7 7.0 Replace Cottonwood 230/115 kV No. 4 XFMR with higher capacity 1 10 10.0 Replace Cottonwood 230/115 kV No. 1 XFMR with higher capacity 1 10 10.0 Voltage Support required at various locations (low voltages throughout entire area) 8 20 160.0 Total Cost for Alternative 944 *Note - Reconductoring costs do not include pole replacements or any substation upgrades

Table 3 – Alternative 2 Associated Work Alternative 2 - Voltage Conversion, Reconductoring and Transformer Upgrades only Description Units P.U. Costs Total Costs Convert Bridgeville - Garberville 60 kV to 115 kV Line 86.4 1.5 129.6 Convert Rio Dell 60 kV Tap 60 kV to 115 kV Line and reconductor 5.8 1.5 8.7 Convert Rio Dell Jct - Bridgeville 60 kV to 115 kV Line and reconductor 21.3 1.5 31.9 Convert Essex Jct. - Arcata - Fairhaven 60 kV Tap 60 kV to 115 kV Line 16.1 1.5 24.1 Convert Fairhaven - Humboldt 60 kV to 115 kV Line 15.3 1.5 22.9 Convert Humboldt Bay - Humboldt No.1 60 kV to 115 kV Line 8.1 1.5 12.1 Convert Humboldt Bay - Rio Dell Jct 60 kV to 115 kV Line 4.4 1.5 6.6 Convert Humboldt Bay - Eureka 60 kV to 115 kV Line 5.6 1.5 8.5 Convert Pacific Lumber (Scotia) 60 kV Tap to 115 kV 5.3 1.5 8.0 Convert Fruitland 60 kV Tap to 115 kV 4.3 1.5 6.4 Convert Fort Seward 60 kV Tap to 115 kV 7.6 1.5 11.5 Convert Humboldt - Maple Creek 60 kV to 115 kV Line 13.9 1.5 20.9 Convert Trinity - Maple Creek 60 kV to 115 kV Line 61.1 1.5 91.7 Convert Garberville - Layonville 60 kV to 115 kV Line 40.0 1.5 60.0 Convert Bridgeville - Cottonwood 115 kV to 230 kV Line 86.4 1.7 146.8 Convert Trinity - Cottonwood 115 kV to 230 kV Line 45.8 1.7 77.9

8

192

Alternative 2 - Voltage Conversion, Reconductoring and Transformer Upgrades only (cont.) Reconductor Willits - Laytonville 60 kV Line 23.4 1.5 35.1 Reconductor Humboldt - Trinity 115 kV Line 68.6 1.5 103.0 Replace Bridgeville 115/60 kV to 230/115 kV XFMR with higher capacity 1 7 7.0 Replace Cottonwood 230/115 kV No. 4 XFMR with higher capacity 1 10 10.0 Replace Cottonwood 230/115 kV No. 1 XFMR with higher capacity 1 10 10.0 Install New Trinity 230/115 kV XFMR 1 7 7.0 Install New Laytonville 115/60 kV XFMR 1 7 7.0 Install New Maple Creek 115 kV XFMR 1 7 7.0 Build New Trinity 230 kV Substation 1 13 13.0 Build New Laytonville 115 kV Substation 1 10.5 10.5 Build New Maple Creek 115 kV Substation 1 10.5 10.5 Convert following 60 kV Substations to 115 kV Substations: Kekawaka 1 4 4.0 Scotia 1 4 4.0 Q0212 1 4 4.0 Eel River 1 4 4.0 Newburg 1 4 4.0 Rio Dell Tap 1 4 4.0 Rio Dell 1 4 4.0 Carlotta 1 4 4.0 Pacific Lumber 1 4 4.0 Swains Flat 1 4 4.0 Fruitland 1 4 4.0 Fort Seward 1 4 4.0 Fairhaven 1 4 4.0 Eureka 1 4 4.0 Maple Creek 1 4 4.0 Arcata 1 4 4.0 Eel River 1 4 4.0 Garberville 1 4 4.0 Convert following 115 kV Substations to 230 kV Substations: Grouse Creek 1 6 6.0 Big Bar 1 6 6.0 Hyampom 1 6 6.0 Ridge Cabin 1 6 6.0 Low Gap 1 6 6.0 Forest Glen 1 6 6.0 Wildwood 1 6 6.0 Total Cost for Alternative 1002

9

193

Table 4 – Alternative 3 Associated Work Alternative 3 - Build New DCTL 230 kV Line and Reconductor only Description Units P.U. Costs Total Costs Build New 230 kV Substation at Rio Dell 1 13 13.0 Reconductor Rio Dell 60 kV Tap 5.8 1.5 8.7 Build New DCTL Rio Dell - Cottonwood 230 kV Line 95 2.4 228.0 SPS 2 5 10.0 Total Cost for Alternative 260

Note: all upgrades are associated with Category B contingencies. Category C violations are mitigated by upgrades acquired for Category B. All costs are based in millions. Costs assume larger insulators, pole replacement, and conductor condition for voltage conversion. Land and permitting not included in this cost estimate.

3. Identify key bottlenecks in the system and provide recommendations on key upgrades that would be most cost-effective in enabling the build-out of future renewable generation. Also, identify geographic areas that are particularly problematic for adding new generation.

Bottlenecks and Recommendations The key bottleneck identified in the Humboldt transmission system is the voltage level (mainly 60 kV) the system is currently operated at and the conductor sizes. The Humboldt transmission system is currently operated as a balanced system of internal generation with some imports and has not been designed to withstand exporting large amounts of power outside of its system. The 60 kV transmission system is not adequate to transport mid to large amounts of generation and would be considered the main limitation or system weakness in converting Humboldt into a generation exporting area. Excess generation in this system results in many normal overloads of the 60 kV system and during contingencies, the same amount of generation needs to be pushed through the rest of the 60 kV circuits causing even more severe overloads. In addition, large flows on low voltage lines cause heavy reactive power losses in the lines which can lead to voltage violations and potential grid voltage instability.

Since Humboldt County is not a densely populated area the amount of generation being proposed would indeed need an outlet onto the main transmission bulk system. Humboldt, being a self sustained system, does not rely heavily on power being imported in therefore the only major tie comes from Cottonwood. This arrangement is reliable and adequate for the current configuration of the system, but does not allow the system to become a generation exporting system. Ideally, as the alternative analysis shows, it would make sense to build new strategically located 230 kV lines that would enable the build-out of renewables and transport of the electric power. This transmission system reinforcement, while difficult to implement appears to be the most efficient and economic way to allow the Humboldt transmission system to export electric power from the proposed renewable generation.

10

194 4. Explain whether or not the various system upgrades would be considered “Network Upgrades” according to CAISO and whether or not the interconnection customer would be repaid the cost of associated Network Upgrades over a five-year period commencing on the Commercial Operation Date of their newly installed generation.

The CAISO Tariff defines Network Upgrades as “The additions, modifications, and upgrades to the CAISO Controlled Grid required at or beyond the Point of Interconnection to accommodate the interconnection of the Generating Facility to the CAISO Controlled Grid. Network Upgrades shall consist of Delivery Network Upgrades and Reliability Network Upgrades. Network Upgrades do not include Distribution Upgrades."

As part of the CAISO Generation Interconnection Procedures (GIP) interconnection studies, Network Upgrades will be identified to mitigate any adverse transmission impacts as a result of the proposed generation facilities. Since the Project has not yet been studied in the CAISO process, it is difficult to determine the full scope of Network Upgrades needed to interconnect the Project. The required upgrades may be the same, more, or less than what is identified in this document. The results are highly dependent on higher-queued projects as well as the other proposed projects that will be studied in the cluster.

The financing of Network Upgrades is the responsibility of each Interconnection Customer up to their cost responsibility as assigned in the interconnection studies. Upon the Commercial Operation Date of the facility, the interconnection customer is entitled to repayment of its contribution to the cost of Network Upgrades. The repayment is made within a five-year period of the Commercial Operation Date and will include interest calculated in accordance with FERC regulations.

11

195